diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,35542 @@ +{ + "best_metric": 0.06506887932492623, + "best_model_checkpoint": "./models/kubhist-byt5-small/checkpoint-48000", + "epoch": 75.18796992481202, + "eval_steps": 1000, + "global_step": 50000, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.015037593984962405, + "grad_norm": 104.62593078613281, + "learning_rate": 5e-07, + "loss": 19.837, + "step": 10 + }, + { + "epoch": 0.03007518796992481, + "grad_norm": 189.65853881835938, + "learning_rate": 1e-06, + "loss": 19.7652, + "step": 20 + }, + { + "epoch": 0.045112781954887216, + "grad_norm": 447.0132751464844, + "learning_rate": 1.5e-06, + "loss": 20.3571, + "step": 30 + }, + { + "epoch": 0.06015037593984962, + "grad_norm": 139.93751525878906, + "learning_rate": 2e-06, + "loss": 20.6364, + "step": 40 + }, + { + "epoch": 0.07518796992481203, + "grad_norm": 499.5147399902344, + "learning_rate": 2.5e-06, + "loss": 20.6889, + "step": 50 + }, + { + "epoch": 0.09022556390977443, + "grad_norm": 478.6932373046875, + "learning_rate": 3e-06, + "loss": 20.6177, + "step": 60 + }, + { + "epoch": 0.10526315789473684, + "grad_norm": 178.10162353515625, + "learning_rate": 3.5e-06, + "loss": 19.9915, + "step": 70 + }, + { + "epoch": 0.12030075187969924, + "grad_norm": 153.1753692626953, + "learning_rate": 4e-06, + "loss": 19.6978, + "step": 80 + }, + { + "epoch": 0.13533834586466165, + "grad_norm": 187.50491333007812, + "learning_rate": 4.5e-06, + "loss": 20.4125, + "step": 90 + }, + { + "epoch": 0.15037593984962405, + "grad_norm": 146.68077087402344, + "learning_rate": 5e-06, + "loss": 20.55, + "step": 100 + }, + { + "epoch": 0.16541353383458646, + "grad_norm": 285.5688781738281, + "learning_rate": 5.5e-06, + "loss": 20.0693, + "step": 110 + }, + { + "epoch": 0.18045112781954886, + "grad_norm": 119.8171157836914, + "learning_rate": 6e-06, + "loss": 19.6952, + "step": 120 + }, + { + "epoch": 0.19548872180451127, + "grad_norm": 137.97911071777344, + "learning_rate": 6.5e-06, + "loss": 18.6694, + "step": 130 + }, + { + "epoch": 0.21052631578947367, + "grad_norm": 316.4723205566406, + "learning_rate": 7e-06, + "loss": 18.8416, + "step": 140 + }, + { + "epoch": 0.22556390977443608, + "grad_norm": 181.7510223388672, + "learning_rate": 7.5e-06, + "loss": 18.8299, + "step": 150 + }, + { + "epoch": 0.24060150375939848, + "grad_norm": 135.28289794921875, + "learning_rate": 8e-06, + "loss": 18.1852, + "step": 160 + }, + { + "epoch": 0.2556390977443609, + "grad_norm": 168.58982849121094, + "learning_rate": 8.500000000000002e-06, + "loss": 17.7585, + "step": 170 + }, + { + "epoch": 0.2706766917293233, + "grad_norm": 136.21595764160156, + "learning_rate": 9e-06, + "loss": 17.4761, + "step": 180 + }, + { + "epoch": 0.2857142857142857, + "grad_norm": 152.5988311767578, + "learning_rate": 9.5e-06, + "loss": 17.6872, + "step": 190 + }, + { + "epoch": 0.3007518796992481, + "grad_norm": 96.53507232666016, + "learning_rate": 1e-05, + "loss": 17.0313, + "step": 200 + }, + { + "epoch": 0.3157894736842105, + "grad_norm": 88.17157745361328, + "learning_rate": 1.0500000000000001e-05, + "loss": 16.3417, + "step": 210 + }, + { + "epoch": 0.3308270676691729, + "grad_norm": 145.6571044921875, + "learning_rate": 1.1e-05, + "loss": 15.5366, + "step": 220 + }, + { + "epoch": 0.3458646616541353, + "grad_norm": 187.11985778808594, + "learning_rate": 1.15e-05, + "loss": 15.6082, + "step": 230 + }, + { + "epoch": 0.3609022556390977, + "grad_norm": 81.12650299072266, + "learning_rate": 1.2e-05, + "loss": 14.5112, + "step": 240 + }, + { + "epoch": 0.37593984962406013, + "grad_norm": 82.21636199951172, + "learning_rate": 1.25e-05, + "loss": 14.9326, + "step": 250 + }, + { + "epoch": 0.39097744360902253, + "grad_norm": 85.29484558105469, + "learning_rate": 1.3e-05, + "loss": 13.9079, + "step": 260 + }, + { + "epoch": 0.40601503759398494, + "grad_norm": 62.190773010253906, + "learning_rate": 1.35e-05, + "loss": 13.7616, + "step": 270 + }, + { + "epoch": 0.42105263157894735, + "grad_norm": 59.5329704284668, + "learning_rate": 1.4e-05, + "loss": 12.6397, + "step": 280 + }, + { + "epoch": 0.43609022556390975, + "grad_norm": 52.637325286865234, + "learning_rate": 1.4500000000000002e-05, + "loss": 11.3343, + "step": 290 + }, + { + "epoch": 0.45112781954887216, + "grad_norm": 61.02895736694336, + "learning_rate": 1.5e-05, + "loss": 11.1318, + "step": 300 + }, + { + "epoch": 0.46616541353383456, + "grad_norm": 45.35968017578125, + "learning_rate": 1.55e-05, + "loss": 10.0478, + "step": 310 + }, + { + "epoch": 0.48120300751879697, + "grad_norm": 52.0111083984375, + "learning_rate": 1.6e-05, + "loss": 9.2311, + "step": 320 + }, + { + "epoch": 0.49624060150375937, + "grad_norm": 53.7806510925293, + "learning_rate": 1.65e-05, + "loss": 8.7852, + "step": 330 + }, + { + "epoch": 0.5112781954887218, + "grad_norm": 144.44883728027344, + "learning_rate": 1.7000000000000003e-05, + "loss": 7.4075, + "step": 340 + }, + { + "epoch": 0.5263157894736842, + "grad_norm": 43.95303726196289, + "learning_rate": 1.7500000000000002e-05, + "loss": 6.0582, + "step": 350 + }, + { + "epoch": 0.5413533834586466, + "grad_norm": 29.293272018432617, + "learning_rate": 1.8e-05, + "loss": 4.6841, + "step": 360 + }, + { + "epoch": 0.556390977443609, + "grad_norm": 27.286880493164062, + "learning_rate": 1.85e-05, + "loss": 3.4778, + "step": 370 + }, + { + "epoch": 0.5714285714285714, + "grad_norm": 13.270530700683594, + "learning_rate": 1.9e-05, + "loss": 2.4978, + "step": 380 + }, + { + "epoch": 0.5864661654135338, + "grad_norm": 5.561248302459717, + "learning_rate": 1.95e-05, + "loss": 2.0326, + "step": 390 + }, + { + "epoch": 0.6015037593984962, + "grad_norm": 4.807607173919678, + "learning_rate": 2e-05, + "loss": 1.8175, + "step": 400 + }, + { + "epoch": 0.6165413533834586, + "grad_norm": 5.500110149383545, + "learning_rate": 2.05e-05, + "loss": 1.7192, + "step": 410 + }, + { + "epoch": 0.631578947368421, + "grad_norm": 4.458729267120361, + "learning_rate": 2.1000000000000002e-05, + "loss": 1.6349, + "step": 420 + }, + { + "epoch": 0.6466165413533834, + "grad_norm": 4.177635192871094, + "learning_rate": 2.1499999999999997e-05, + "loss": 1.6121, + "step": 430 + }, + { + "epoch": 0.6616541353383458, + "grad_norm": 4.807072639465332, + "learning_rate": 2.2e-05, + "loss": 1.5495, + "step": 440 + }, + { + "epoch": 0.6766917293233082, + "grad_norm": 4.354175567626953, + "learning_rate": 2.2499999999999998e-05, + "loss": 1.5219, + "step": 450 + }, + { + "epoch": 0.6917293233082706, + "grad_norm": 5.061458587646484, + "learning_rate": 2.3e-05, + "loss": 1.4078, + "step": 460 + }, + { + "epoch": 0.706766917293233, + "grad_norm": 6.091569900512695, + "learning_rate": 2.3500000000000002e-05, + "loss": 1.3904, + "step": 470 + }, + { + "epoch": 0.7218045112781954, + "grad_norm": 5.267693996429443, + "learning_rate": 2.4e-05, + "loss": 1.3134, + "step": 480 + }, + { + "epoch": 0.7368421052631579, + "grad_norm": 6.588474273681641, + "learning_rate": 2.4500000000000003e-05, + "loss": 1.2044, + "step": 490 + }, + { + "epoch": 0.7518796992481203, + "grad_norm": 6.151535987854004, + "learning_rate": 2.5e-05, + "loss": 1.0959, + "step": 500 + }, + { + "epoch": 0.7669172932330827, + "grad_norm": 6.021274566650391, + "learning_rate": 2.55e-05, + "loss": 0.9294, + "step": 510 + }, + { + "epoch": 0.7819548872180451, + "grad_norm": 7.281777858734131, + "learning_rate": 2.6e-05, + "loss": 0.8163, + "step": 520 + }, + { + "epoch": 0.7969924812030075, + "grad_norm": 6.973840236663818, + "learning_rate": 2.65e-05, + "loss": 0.7043, + "step": 530 + }, + { + "epoch": 0.8120300751879699, + "grad_norm": 6.8748979568481445, + "learning_rate": 2.7e-05, + "loss": 0.6174, + "step": 540 + }, + { + "epoch": 0.8270676691729323, + "grad_norm": 6.0368876457214355, + "learning_rate": 2.75e-05, + "loss": 0.5241, + "step": 550 + }, + { + "epoch": 0.8421052631578947, + "grad_norm": 7.205780982971191, + "learning_rate": 2.8e-05, + "loss": 0.48, + "step": 560 + }, + { + "epoch": 0.8571428571428571, + "grad_norm": 5.368173122406006, + "learning_rate": 2.85e-05, + "loss": 0.404, + "step": 570 + }, + { + "epoch": 0.8721804511278195, + "grad_norm": 5.182989120483398, + "learning_rate": 2.9000000000000004e-05, + "loss": 0.3454, + "step": 580 + }, + { + "epoch": 0.8872180451127819, + "grad_norm": 4.2408671379089355, + "learning_rate": 2.95e-05, + "loss": 0.2956, + "step": 590 + }, + { + "epoch": 0.9022556390977443, + "grad_norm": 3.6516077518463135, + "learning_rate": 3e-05, + "loss": 0.2621, + "step": 600 + }, + { + "epoch": 0.9172932330827067, + "grad_norm": 2.526935338973999, + "learning_rate": 3.05e-05, + "loss": 0.2522, + "step": 610 + }, + { + "epoch": 0.9323308270676691, + "grad_norm": 1.9546092748641968, + "learning_rate": 3.1e-05, + "loss": 0.218, + "step": 620 + }, + { + "epoch": 0.9473684210526315, + "grad_norm": 1.1249884366989136, + "learning_rate": 3.15e-05, + "loss": 0.1948, + "step": 630 + }, + { + "epoch": 0.9624060150375939, + "grad_norm": 1.7423558235168457, + "learning_rate": 3.2e-05, + "loss": 0.1776, + "step": 640 + }, + { + "epoch": 0.9774436090225563, + "grad_norm": 0.9116438031196594, + "learning_rate": 3.2500000000000004e-05, + "loss": 0.1675, + "step": 650 + }, + { + "epoch": 0.9924812030075187, + "grad_norm": 0.6737796068191528, + "learning_rate": 3.3e-05, + "loss": 0.1623, + "step": 660 + }, + { + "epoch": 1.0075187969924813, + "grad_norm": 0.5420631766319275, + "learning_rate": 3.35e-05, + "loss": 0.1476, + "step": 670 + }, + { + "epoch": 1.0225563909774436, + "grad_norm": 0.41047796607017517, + "learning_rate": 3.4000000000000007e-05, + "loss": 0.1499, + "step": 680 + }, + { + "epoch": 1.037593984962406, + "grad_norm": 0.3725735545158386, + "learning_rate": 3.4500000000000005e-05, + "loss": 0.1343, + "step": 690 + }, + { + "epoch": 1.0526315789473684, + "grad_norm": 0.38603830337524414, + "learning_rate": 3.5000000000000004e-05, + "loss": 0.1483, + "step": 700 + }, + { + "epoch": 1.0676691729323309, + "grad_norm": 0.3030599355697632, + "learning_rate": 3.5499999999999996e-05, + "loss": 0.1375, + "step": 710 + }, + { + "epoch": 1.0827067669172932, + "grad_norm": 0.5150690078735352, + "learning_rate": 3.6e-05, + "loss": 0.138, + "step": 720 + }, + { + "epoch": 1.0977443609022557, + "grad_norm": 0.29648077487945557, + "learning_rate": 3.65e-05, + "loss": 0.127, + "step": 730 + }, + { + "epoch": 1.112781954887218, + "grad_norm": 0.33018654584884644, + "learning_rate": 3.7e-05, + "loss": 0.1257, + "step": 740 + }, + { + "epoch": 1.1278195488721805, + "grad_norm": 0.45184335112571716, + "learning_rate": 3.75e-05, + "loss": 0.12, + "step": 750 + }, + { + "epoch": 1.1428571428571428, + "grad_norm": 0.23905764520168304, + "learning_rate": 3.8e-05, + "loss": 0.1196, + "step": 760 + }, + { + "epoch": 1.1578947368421053, + "grad_norm": 0.20826616883277893, + "learning_rate": 3.85e-05, + "loss": 0.1271, + "step": 770 + }, + { + "epoch": 1.1729323308270676, + "grad_norm": 0.22215485572814941, + "learning_rate": 3.9e-05, + "loss": 0.1185, + "step": 780 + }, + { + "epoch": 1.1879699248120301, + "grad_norm": 0.22439247369766235, + "learning_rate": 3.95e-05, + "loss": 0.1206, + "step": 790 + }, + { + "epoch": 1.2030075187969924, + "grad_norm": 0.22341278195381165, + "learning_rate": 4e-05, + "loss": 0.1149, + "step": 800 + }, + { + "epoch": 1.218045112781955, + "grad_norm": 0.23636086285114288, + "learning_rate": 4.05e-05, + "loss": 0.1145, + "step": 810 + }, + { + "epoch": 1.2330827067669172, + "grad_norm": 0.1792980134487152, + "learning_rate": 4.1e-05, + "loss": 0.119, + "step": 820 + }, + { + "epoch": 1.2481203007518797, + "grad_norm": 0.21891237795352936, + "learning_rate": 4.1500000000000006e-05, + "loss": 0.1164, + "step": 830 + }, + { + "epoch": 1.263157894736842, + "grad_norm": 0.18507832288742065, + "learning_rate": 4.2000000000000004e-05, + "loss": 0.1178, + "step": 840 + }, + { + "epoch": 1.2781954887218046, + "grad_norm": 0.30749213695526123, + "learning_rate": 4.25e-05, + "loss": 0.1014, + "step": 850 + }, + { + "epoch": 1.2932330827067668, + "grad_norm": 0.22607159614562988, + "learning_rate": 4.2999999999999995e-05, + "loss": 0.1111, + "step": 860 + }, + { + "epoch": 1.3082706766917294, + "grad_norm": 0.15399128198623657, + "learning_rate": 4.35e-05, + "loss": 0.1046, + "step": 870 + }, + { + "epoch": 1.3233082706766917, + "grad_norm": 0.22664740681648254, + "learning_rate": 4.4e-05, + "loss": 0.1044, + "step": 880 + }, + { + "epoch": 1.3383458646616542, + "grad_norm": 0.1503833681344986, + "learning_rate": 4.45e-05, + "loss": 0.1074, + "step": 890 + }, + { + "epoch": 1.3533834586466165, + "grad_norm": 0.15727737545967102, + "learning_rate": 4.4999999999999996e-05, + "loss": 0.1019, + "step": 900 + }, + { + "epoch": 1.368421052631579, + "grad_norm": 0.1947244256734848, + "learning_rate": 4.55e-05, + "loss": 0.1008, + "step": 910 + }, + { + "epoch": 1.3834586466165413, + "grad_norm": 0.14236390590667725, + "learning_rate": 4.6e-05, + "loss": 0.1019, + "step": 920 + }, + { + "epoch": 1.3984962406015038, + "grad_norm": 0.32630428671836853, + "learning_rate": 4.65e-05, + "loss": 0.1019, + "step": 930 + }, + { + "epoch": 1.413533834586466, + "grad_norm": 0.13208597898483276, + "learning_rate": 4.7000000000000004e-05, + "loss": 0.1041, + "step": 940 + }, + { + "epoch": 1.4285714285714286, + "grad_norm": 0.13336873054504395, + "learning_rate": 4.75e-05, + "loss": 0.1046, + "step": 950 + }, + { + "epoch": 1.443609022556391, + "grad_norm": 0.15105360746383667, + "learning_rate": 4.8e-05, + "loss": 0.1029, + "step": 960 + }, + { + "epoch": 1.4586466165413534, + "grad_norm": 0.18850146234035492, + "learning_rate": 4.85e-05, + "loss": 0.1055, + "step": 970 + }, + { + "epoch": 1.4736842105263157, + "grad_norm": 0.12361064553260803, + "learning_rate": 4.9000000000000005e-05, + "loss": 0.0996, + "step": 980 + }, + { + "epoch": 1.4887218045112782, + "grad_norm": 0.17332784831523895, + "learning_rate": 4.9500000000000004e-05, + "loss": 0.1003, + "step": 990 + }, + { + "epoch": 1.5037593984962405, + "grad_norm": 0.12834808230400085, + "learning_rate": 5e-05, + "loss": 0.1047, + "step": 1000 + }, + { + "epoch": 1.5037593984962405, + "eval_cer": 0.02805686314246116, + "eval_loss": 0.0826926901936531, + "eval_runtime": 156.545, + "eval_samples_per_second": 102.38, + "eval_steps_per_second": 0.805, + "eval_wer": 0.11151308989080795, + "step": 1000 + }, + { + "epoch": 1.518796992481203, + "grad_norm": 0.12339483201503754, + "learning_rate": 5.05e-05, + "loss": 0.096, + "step": 1010 + }, + { + "epoch": 1.5338345864661656, + "grad_norm": 0.1696172058582306, + "learning_rate": 5.1e-05, + "loss": 0.0953, + "step": 1020 + }, + { + "epoch": 1.5488721804511278, + "grad_norm": 0.14241768419742584, + "learning_rate": 5.15e-05, + "loss": 0.0866, + "step": 1030 + }, + { + "epoch": 1.5639097744360901, + "grad_norm": 0.14582350850105286, + "learning_rate": 5.2e-05, + "loss": 0.0927, + "step": 1040 + }, + { + "epoch": 1.5789473684210527, + "grad_norm": 0.12379587441682816, + "learning_rate": 5.25e-05, + "loss": 0.0944, + "step": 1050 + }, + { + "epoch": 1.5939849624060152, + "grad_norm": 0.14266642928123474, + "learning_rate": 5.3e-05, + "loss": 0.0918, + "step": 1060 + }, + { + "epoch": 1.6090225563909775, + "grad_norm": 0.17819897830486298, + "learning_rate": 5.35e-05, + "loss": 0.0894, + "step": 1070 + }, + { + "epoch": 1.6240601503759398, + "grad_norm": 0.15033525228500366, + "learning_rate": 5.4e-05, + "loss": 0.089, + "step": 1080 + }, + { + "epoch": 1.6390977443609023, + "grad_norm": 0.14557458460330963, + "learning_rate": 5.45e-05, + "loss": 0.0922, + "step": 1090 + }, + { + "epoch": 1.6541353383458648, + "grad_norm": 0.14062285423278809, + "learning_rate": 5.5e-05, + "loss": 0.0932, + "step": 1100 + }, + { + "epoch": 1.669172932330827, + "grad_norm": 0.14711818099021912, + "learning_rate": 5.55e-05, + "loss": 0.0936, + "step": 1110 + }, + { + "epoch": 1.6842105263157894, + "grad_norm": 0.1632707566022873, + "learning_rate": 5.6e-05, + "loss": 0.0948, + "step": 1120 + }, + { + "epoch": 1.699248120300752, + "grad_norm": 0.09498897194862366, + "learning_rate": 5.6500000000000005e-05, + "loss": 0.0843, + "step": 1130 + }, + { + "epoch": 1.7142857142857144, + "grad_norm": 0.16259658336639404, + "learning_rate": 5.7e-05, + "loss": 0.0912, + "step": 1140 + }, + { + "epoch": 1.7293233082706767, + "grad_norm": 0.11321844905614853, + "learning_rate": 5.75e-05, + "loss": 0.1003, + "step": 1150 + }, + { + "epoch": 1.744360902255639, + "grad_norm": 0.10526232421398163, + "learning_rate": 5.800000000000001e-05, + "loss": 0.0885, + "step": 1160 + }, + { + "epoch": 1.7593984962406015, + "grad_norm": 0.1256016045808792, + "learning_rate": 5.8500000000000006e-05, + "loss": 0.0858, + "step": 1170 + }, + { + "epoch": 1.774436090225564, + "grad_norm": 0.13412119448184967, + "learning_rate": 5.9e-05, + "loss": 0.092, + "step": 1180 + }, + { + "epoch": 1.7894736842105263, + "grad_norm": 0.4521901309490204, + "learning_rate": 5.9499999999999996e-05, + "loss": 0.0897, + "step": 1190 + }, + { + "epoch": 1.8045112781954886, + "grad_norm": 0.0998324379324913, + "learning_rate": 6e-05, + "loss": 0.0847, + "step": 1200 + }, + { + "epoch": 1.8195488721804511, + "grad_norm": 0.08651789277791977, + "learning_rate": 6.05e-05, + "loss": 0.0887, + "step": 1210 + }, + { + "epoch": 1.8345864661654137, + "grad_norm": 0.1047699898481369, + "learning_rate": 6.1e-05, + "loss": 0.0885, + "step": 1220 + }, + { + "epoch": 1.849624060150376, + "grad_norm": 0.18981552124023438, + "learning_rate": 6.15e-05, + "loss": 0.0866, + "step": 1230 + }, + { + "epoch": 1.8646616541353382, + "grad_norm": 0.10824630409479141, + "learning_rate": 6.2e-05, + "loss": 0.0809, + "step": 1240 + }, + { + "epoch": 1.8796992481203008, + "grad_norm": 0.15814915299415588, + "learning_rate": 6.25e-05, + "loss": 0.0836, + "step": 1250 + }, + { + "epoch": 1.8947368421052633, + "grad_norm": 0.11502755433320999, + "learning_rate": 6.3e-05, + "loss": 0.0837, + "step": 1260 + }, + { + "epoch": 1.9097744360902256, + "grad_norm": 0.09850893169641495, + "learning_rate": 6.35e-05, + "loss": 0.0871, + "step": 1270 + }, + { + "epoch": 1.9248120300751879, + "grad_norm": 0.08723190426826477, + "learning_rate": 6.4e-05, + "loss": 0.0873, + "step": 1280 + }, + { + "epoch": 1.9398496240601504, + "grad_norm": 0.12907691299915314, + "learning_rate": 6.450000000000001e-05, + "loss": 0.0843, + "step": 1290 + }, + { + "epoch": 1.954887218045113, + "grad_norm": 0.08867259323596954, + "learning_rate": 6.500000000000001e-05, + "loss": 0.0824, + "step": 1300 + }, + { + "epoch": 1.9699248120300752, + "grad_norm": 0.08052284270524979, + "learning_rate": 6.55e-05, + "loss": 0.0804, + "step": 1310 + }, + { + "epoch": 1.9849624060150375, + "grad_norm": 0.1067063957452774, + "learning_rate": 6.6e-05, + "loss": 0.0855, + "step": 1320 + }, + { + "epoch": 2.0, + "grad_norm": 0.09426485002040863, + "learning_rate": 6.65e-05, + "loss": 0.0839, + "step": 1330 + }, + { + "epoch": 2.0150375939849625, + "grad_norm": 0.6457942128181458, + "learning_rate": 6.7e-05, + "loss": 0.0807, + "step": 1340 + }, + { + "epoch": 2.030075187969925, + "grad_norm": 0.11477750539779663, + "learning_rate": 6.75e-05, + "loss": 0.0844, + "step": 1350 + }, + { + "epoch": 2.045112781954887, + "grad_norm": 0.09924060851335526, + "learning_rate": 6.800000000000001e-05, + "loss": 0.0882, + "step": 1360 + }, + { + "epoch": 2.0601503759398496, + "grad_norm": 0.15235508978366852, + "learning_rate": 6.850000000000001e-05, + "loss": 0.0751, + "step": 1370 + }, + { + "epoch": 2.075187969924812, + "grad_norm": 0.09801047295331955, + "learning_rate": 6.900000000000001e-05, + "loss": 0.0762, + "step": 1380 + }, + { + "epoch": 2.090225563909774, + "grad_norm": 0.12657718360424042, + "learning_rate": 6.950000000000001e-05, + "loss": 0.0741, + "step": 1390 + }, + { + "epoch": 2.1052631578947367, + "grad_norm": 0.10327859967947006, + "learning_rate": 7.000000000000001e-05, + "loss": 0.0741, + "step": 1400 + }, + { + "epoch": 2.1203007518796992, + "grad_norm": 0.085964135825634, + "learning_rate": 7.049999999999999e-05, + "loss": 0.0771, + "step": 1410 + }, + { + "epoch": 2.1353383458646618, + "grad_norm": 0.11042708158493042, + "learning_rate": 7.099999999999999e-05, + "loss": 0.0764, + "step": 1420 + }, + { + "epoch": 2.1503759398496243, + "grad_norm": 0.08443091809749603, + "learning_rate": 7.149999999999999e-05, + "loss": 0.0766, + "step": 1430 + }, + { + "epoch": 2.1654135338345863, + "grad_norm": 0.08230534940958023, + "learning_rate": 7.2e-05, + "loss": 0.07, + "step": 1440 + }, + { + "epoch": 2.180451127819549, + "grad_norm": 0.10639423131942749, + "learning_rate": 7.25e-05, + "loss": 0.0806, + "step": 1450 + }, + { + "epoch": 2.1954887218045114, + "grad_norm": 0.09049826860427856, + "learning_rate": 7.3e-05, + "loss": 0.0749, + "step": 1460 + }, + { + "epoch": 2.2105263157894735, + "grad_norm": 0.14299000799655914, + "learning_rate": 7.35e-05, + "loss": 0.0753, + "step": 1470 + }, + { + "epoch": 2.225563909774436, + "grad_norm": 0.1076231300830841, + "learning_rate": 7.4e-05, + "loss": 0.0807, + "step": 1480 + }, + { + "epoch": 2.2406015037593985, + "grad_norm": 0.09299056231975555, + "learning_rate": 7.45e-05, + "loss": 0.0778, + "step": 1490 + }, + { + "epoch": 2.255639097744361, + "grad_norm": 0.12077386677265167, + "learning_rate": 7.5e-05, + "loss": 0.0812, + "step": 1500 + }, + { + "epoch": 2.2706766917293235, + "grad_norm": 0.08236552774906158, + "learning_rate": 7.55e-05, + "loss": 0.0779, + "step": 1510 + }, + { + "epoch": 2.2857142857142856, + "grad_norm": 0.1611240804195404, + "learning_rate": 7.6e-05, + "loss": 0.0756, + "step": 1520 + }, + { + "epoch": 2.300751879699248, + "grad_norm": 0.12710203230381012, + "learning_rate": 7.65e-05, + "loss": 0.0702, + "step": 1530 + }, + { + "epoch": 2.3157894736842106, + "grad_norm": 0.09004311263561249, + "learning_rate": 7.7e-05, + "loss": 0.0797, + "step": 1540 + }, + { + "epoch": 2.3308270676691727, + "grad_norm": 0.122159942984581, + "learning_rate": 7.75e-05, + "loss": 0.0764, + "step": 1550 + }, + { + "epoch": 2.345864661654135, + "grad_norm": 0.09201399236917496, + "learning_rate": 7.8e-05, + "loss": 0.0741, + "step": 1560 + }, + { + "epoch": 2.3609022556390977, + "grad_norm": 0.09120766073465347, + "learning_rate": 7.85e-05, + "loss": 0.0733, + "step": 1570 + }, + { + "epoch": 2.3759398496240602, + "grad_norm": 0.10722893476486206, + "learning_rate": 7.9e-05, + "loss": 0.069, + "step": 1580 + }, + { + "epoch": 2.3909774436090228, + "grad_norm": 0.08648216724395752, + "learning_rate": 7.950000000000001e-05, + "loss": 0.0726, + "step": 1590 + }, + { + "epoch": 2.406015037593985, + "grad_norm": 0.17359693348407745, + "learning_rate": 8e-05, + "loss": 0.079, + "step": 1600 + }, + { + "epoch": 2.4210526315789473, + "grad_norm": 0.09163235127925873, + "learning_rate": 8.05e-05, + "loss": 0.074, + "step": 1610 + }, + { + "epoch": 2.43609022556391, + "grad_norm": 0.1336594671010971, + "learning_rate": 8.1e-05, + "loss": 0.0759, + "step": 1620 + }, + { + "epoch": 2.451127819548872, + "grad_norm": 0.0982159748673439, + "learning_rate": 8.15e-05, + "loss": 0.0754, + "step": 1630 + }, + { + "epoch": 2.4661654135338344, + "grad_norm": 0.0917542576789856, + "learning_rate": 8.2e-05, + "loss": 0.0763, + "step": 1640 + }, + { + "epoch": 2.481203007518797, + "grad_norm": 0.08298458158969879, + "learning_rate": 8.25e-05, + "loss": 0.0788, + "step": 1650 + }, + { + "epoch": 2.4962406015037595, + "grad_norm": 0.1058349758386612, + "learning_rate": 8.300000000000001e-05, + "loss": 0.0764, + "step": 1660 + }, + { + "epoch": 2.511278195488722, + "grad_norm": 0.08925671130418777, + "learning_rate": 8.350000000000001e-05, + "loss": 0.0798, + "step": 1670 + }, + { + "epoch": 2.526315789473684, + "grad_norm": 0.0884801596403122, + "learning_rate": 8.400000000000001e-05, + "loss": 0.0705, + "step": 1680 + }, + { + "epoch": 2.5413533834586466, + "grad_norm": 0.1008792594075203, + "learning_rate": 8.450000000000001e-05, + "loss": 0.0678, + "step": 1690 + }, + { + "epoch": 2.556390977443609, + "grad_norm": 0.12110363692045212, + "learning_rate": 8.5e-05, + "loss": 0.0764, + "step": 1700 + }, + { + "epoch": 2.571428571428571, + "grad_norm": 0.10572750866413116, + "learning_rate": 8.55e-05, + "loss": 0.0751, + "step": 1710 + }, + { + "epoch": 2.5864661654135337, + "grad_norm": 0.09433761239051819, + "learning_rate": 8.599999999999999e-05, + "loss": 0.0752, + "step": 1720 + }, + { + "epoch": 2.601503759398496, + "grad_norm": 0.09577427059412003, + "learning_rate": 8.65e-05, + "loss": 0.0718, + "step": 1730 + }, + { + "epoch": 2.6165413533834587, + "grad_norm": 0.08834993839263916, + "learning_rate": 8.7e-05, + "loss": 0.0768, + "step": 1740 + }, + { + "epoch": 2.6315789473684212, + "grad_norm": 0.09512394666671753, + "learning_rate": 8.75e-05, + "loss": 0.0664, + "step": 1750 + }, + { + "epoch": 2.6466165413533833, + "grad_norm": 0.09866785258054733, + "learning_rate": 8.8e-05, + "loss": 0.0756, + "step": 1760 + }, + { + "epoch": 2.661654135338346, + "grad_norm": 0.08628912270069122, + "learning_rate": 8.85e-05, + "loss": 0.069, + "step": 1770 + }, + { + "epoch": 2.6766917293233083, + "grad_norm": 0.09126464277505875, + "learning_rate": 8.9e-05, + "loss": 0.0764, + "step": 1780 + }, + { + "epoch": 2.6917293233082704, + "grad_norm": 0.08471488207578659, + "learning_rate": 8.95e-05, + "loss": 0.0719, + "step": 1790 + }, + { + "epoch": 2.706766917293233, + "grad_norm": 0.074613556265831, + "learning_rate": 8.999999999999999e-05, + "loss": 0.0665, + "step": 1800 + }, + { + "epoch": 2.7218045112781954, + "grad_norm": 0.09572897106409073, + "learning_rate": 9.05e-05, + "loss": 0.0734, + "step": 1810 + }, + { + "epoch": 2.736842105263158, + "grad_norm": 0.07093536853790283, + "learning_rate": 9.1e-05, + "loss": 0.069, + "step": 1820 + }, + { + "epoch": 2.7518796992481205, + "grad_norm": 0.08359956741333008, + "learning_rate": 9.15e-05, + "loss": 0.0692, + "step": 1830 + }, + { + "epoch": 2.7669172932330826, + "grad_norm": 0.14625804126262665, + "learning_rate": 9.2e-05, + "loss": 0.0659, + "step": 1840 + }, + { + "epoch": 2.781954887218045, + "grad_norm": 0.08347858488559723, + "learning_rate": 9.25e-05, + "loss": 0.0669, + "step": 1850 + }, + { + "epoch": 2.7969924812030076, + "grad_norm": 0.09789352118968964, + "learning_rate": 9.3e-05, + "loss": 0.0641, + "step": 1860 + }, + { + "epoch": 2.8120300751879697, + "grad_norm": 0.07653295993804932, + "learning_rate": 9.35e-05, + "loss": 0.0682, + "step": 1870 + }, + { + "epoch": 2.827067669172932, + "grad_norm": 0.08785437047481537, + "learning_rate": 9.400000000000001e-05, + "loss": 0.0641, + "step": 1880 + }, + { + "epoch": 2.8421052631578947, + "grad_norm": 0.0775570347905159, + "learning_rate": 9.45e-05, + "loss": 0.0707, + "step": 1890 + }, + { + "epoch": 2.857142857142857, + "grad_norm": 0.0973789319396019, + "learning_rate": 9.5e-05, + "loss": 0.0711, + "step": 1900 + }, + { + "epoch": 2.8721804511278197, + "grad_norm": 0.08315828442573547, + "learning_rate": 9.55e-05, + "loss": 0.066, + "step": 1910 + }, + { + "epoch": 2.887218045112782, + "grad_norm": 0.08287108689546585, + "learning_rate": 9.6e-05, + "loss": 0.0664, + "step": 1920 + }, + { + "epoch": 2.9022556390977443, + "grad_norm": 0.0989585667848587, + "learning_rate": 9.65e-05, + "loss": 0.0681, + "step": 1930 + }, + { + "epoch": 2.917293233082707, + "grad_norm": 0.09089252352714539, + "learning_rate": 9.7e-05, + "loss": 0.0675, + "step": 1940 + }, + { + "epoch": 2.932330827067669, + "grad_norm": 0.09598240256309509, + "learning_rate": 9.750000000000001e-05, + "loss": 0.0722, + "step": 1950 + }, + { + "epoch": 2.9473684210526314, + "grad_norm": 0.07312577217817307, + "learning_rate": 9.800000000000001e-05, + "loss": 0.0676, + "step": 1960 + }, + { + "epoch": 2.962406015037594, + "grad_norm": 0.1499778926372528, + "learning_rate": 9.850000000000001e-05, + "loss": 0.0651, + "step": 1970 + }, + { + "epoch": 2.9774436090225564, + "grad_norm": 0.10445220023393631, + "learning_rate": 9.900000000000001e-05, + "loss": 0.065, + "step": 1980 + }, + { + "epoch": 2.992481203007519, + "grad_norm": 0.08579041063785553, + "learning_rate": 9.95e-05, + "loss": 0.0637, + "step": 1990 + }, + { + "epoch": 3.007518796992481, + "grad_norm": 0.07390273362398148, + "learning_rate": 0.0001, + "loss": 0.0683, + "step": 2000 + }, + { + "epoch": 3.007518796992481, + "eval_cer": 0.02314621504027942, + "eval_loss": 0.060988396406173706, + "eval_runtime": 157.9295, + "eval_samples_per_second": 101.482, + "eval_steps_per_second": 0.798, + "eval_wer": 0.08864101937642128, + "step": 2000 + }, + { + "epoch": 3.0225563909774436, + "grad_norm": 0.09385969489812851, + "learning_rate": 0.0001005, + "loss": 0.064, + "step": 2010 + }, + { + "epoch": 3.037593984962406, + "grad_norm": 0.09060752391815186, + "learning_rate": 0.000101, + "loss": 0.0675, + "step": 2020 + }, + { + "epoch": 3.0526315789473686, + "grad_norm": 0.10236264765262604, + "learning_rate": 0.00010150000000000001, + "loss": 0.0673, + "step": 2030 + }, + { + "epoch": 3.0676691729323307, + "grad_norm": 0.07667925208806992, + "learning_rate": 0.000102, + "loss": 0.0602, + "step": 2040 + }, + { + "epoch": 3.082706766917293, + "grad_norm": 0.07684729248285294, + "learning_rate": 0.0001025, + "loss": 0.0617, + "step": 2050 + }, + { + "epoch": 3.0977443609022557, + "grad_norm": 0.08042387664318085, + "learning_rate": 0.000103, + "loss": 0.0604, + "step": 2060 + }, + { + "epoch": 3.112781954887218, + "grad_norm": 0.08780447393655777, + "learning_rate": 0.0001035, + "loss": 0.0628, + "step": 2070 + }, + { + "epoch": 3.1278195488721803, + "grad_norm": 0.06573078036308289, + "learning_rate": 0.000104, + "loss": 0.0638, + "step": 2080 + }, + { + "epoch": 3.142857142857143, + "grad_norm": 0.08158091455698013, + "learning_rate": 0.00010449999999999999, + "loss": 0.0652, + "step": 2090 + }, + { + "epoch": 3.1578947368421053, + "grad_norm": 0.07186536490917206, + "learning_rate": 0.000105, + "loss": 0.0656, + "step": 2100 + }, + { + "epoch": 3.172932330827068, + "grad_norm": 0.07477926462888718, + "learning_rate": 0.0001055, + "loss": 0.0668, + "step": 2110 + }, + { + "epoch": 3.18796992481203, + "grad_norm": 0.08838914334774017, + "learning_rate": 0.000106, + "loss": 0.0612, + "step": 2120 + }, + { + "epoch": 3.2030075187969924, + "grad_norm": 0.06792841106653214, + "learning_rate": 0.0001065, + "loss": 0.064, + "step": 2130 + }, + { + "epoch": 3.218045112781955, + "grad_norm": 0.08047439903020859, + "learning_rate": 0.000107, + "loss": 0.0672, + "step": 2140 + }, + { + "epoch": 3.2330827067669174, + "grad_norm": 0.12146124243736267, + "learning_rate": 0.0001075, + "loss": 0.0672, + "step": 2150 + }, + { + "epoch": 3.2481203007518795, + "grad_norm": 0.0637350007891655, + "learning_rate": 0.000108, + "loss": 0.06, + "step": 2160 + }, + { + "epoch": 3.263157894736842, + "grad_norm": 0.07478287070989609, + "learning_rate": 0.00010850000000000001, + "loss": 0.0656, + "step": 2170 + }, + { + "epoch": 3.2781954887218046, + "grad_norm": 0.08031600713729858, + "learning_rate": 0.000109, + "loss": 0.0633, + "step": 2180 + }, + { + "epoch": 3.293233082706767, + "grad_norm": 0.204091414809227, + "learning_rate": 0.0001095, + "loss": 0.0589, + "step": 2190 + }, + { + "epoch": 3.308270676691729, + "grad_norm": 0.08402310311794281, + "learning_rate": 0.00011, + "loss": 0.0591, + "step": 2200 + }, + { + "epoch": 3.3233082706766917, + "grad_norm": 0.07302872836589813, + "learning_rate": 0.0001105, + "loss": 0.0626, + "step": 2210 + }, + { + "epoch": 3.338345864661654, + "grad_norm": 0.07210084795951843, + "learning_rate": 0.000111, + "loss": 0.0627, + "step": 2220 + }, + { + "epoch": 3.3533834586466167, + "grad_norm": 0.11313608288764954, + "learning_rate": 0.0001115, + "loss": 0.0653, + "step": 2230 + }, + { + "epoch": 3.3684210526315788, + "grad_norm": 0.0727371796965599, + "learning_rate": 0.000112, + "loss": 0.0599, + "step": 2240 + }, + { + "epoch": 3.3834586466165413, + "grad_norm": 0.0675148293375969, + "learning_rate": 0.00011250000000000001, + "loss": 0.0626, + "step": 2250 + }, + { + "epoch": 3.398496240601504, + "grad_norm": 0.07077009975910187, + "learning_rate": 0.00011300000000000001, + "loss": 0.0615, + "step": 2260 + }, + { + "epoch": 3.4135338345864663, + "grad_norm": 0.07612776756286621, + "learning_rate": 0.00011350000000000001, + "loss": 0.0608, + "step": 2270 + }, + { + "epoch": 3.4285714285714284, + "grad_norm": 0.07042662799358368, + "learning_rate": 0.000114, + "loss": 0.0556, + "step": 2280 + }, + { + "epoch": 3.443609022556391, + "grad_norm": 0.10366267710924149, + "learning_rate": 0.0001145, + "loss": 0.0616, + "step": 2290 + }, + { + "epoch": 3.4586466165413534, + "grad_norm": 0.06444589048624039, + "learning_rate": 0.000115, + "loss": 0.0578, + "step": 2300 + }, + { + "epoch": 3.473684210526316, + "grad_norm": 0.08169475942850113, + "learning_rate": 0.0001155, + "loss": 0.0616, + "step": 2310 + }, + { + "epoch": 3.488721804511278, + "grad_norm": 1.8280595541000366, + "learning_rate": 0.00011600000000000001, + "loss": 0.0601, + "step": 2320 + }, + { + "epoch": 3.5037593984962405, + "grad_norm": 0.07769350707530975, + "learning_rate": 0.00011650000000000001, + "loss": 0.0578, + "step": 2330 + }, + { + "epoch": 3.518796992481203, + "grad_norm": 0.11786318570375443, + "learning_rate": 0.00011700000000000001, + "loss": 0.0583, + "step": 2340 + }, + { + "epoch": 3.5338345864661656, + "grad_norm": 0.0765058845281601, + "learning_rate": 0.0001175, + "loss": 0.0592, + "step": 2350 + }, + { + "epoch": 3.548872180451128, + "grad_norm": 0.07647982984781265, + "learning_rate": 0.000118, + "loss": 0.0541, + "step": 2360 + }, + { + "epoch": 3.56390977443609, + "grad_norm": 0.05874662101268768, + "learning_rate": 0.0001185, + "loss": 0.0599, + "step": 2370 + }, + { + "epoch": 3.5789473684210527, + "grad_norm": 0.0920197144150734, + "learning_rate": 0.00011899999999999999, + "loss": 0.0618, + "step": 2380 + }, + { + "epoch": 3.593984962406015, + "grad_norm": 0.08722665160894394, + "learning_rate": 0.00011949999999999999, + "loss": 0.0614, + "step": 2390 + }, + { + "epoch": 3.6090225563909772, + "grad_norm": 0.06825357675552368, + "learning_rate": 0.00012, + "loss": 0.0598, + "step": 2400 + }, + { + "epoch": 3.6240601503759398, + "grad_norm": 0.062879778444767, + "learning_rate": 0.0001205, + "loss": 0.0594, + "step": 2410 + }, + { + "epoch": 3.6390977443609023, + "grad_norm": 0.08360376954078674, + "learning_rate": 0.000121, + "loss": 0.0609, + "step": 2420 + }, + { + "epoch": 3.654135338345865, + "grad_norm": 0.0683365911245346, + "learning_rate": 0.0001215, + "loss": 0.0566, + "step": 2430 + }, + { + "epoch": 3.6691729323308273, + "grad_norm": 0.10033878684043884, + "learning_rate": 0.000122, + "loss": 0.0593, + "step": 2440 + }, + { + "epoch": 3.6842105263157894, + "grad_norm": 0.08435523509979248, + "learning_rate": 0.0001225, + "loss": 0.0593, + "step": 2450 + }, + { + "epoch": 3.699248120300752, + "grad_norm": 0.0788862332701683, + "learning_rate": 0.000123, + "loss": 0.0602, + "step": 2460 + }, + { + "epoch": 3.7142857142857144, + "grad_norm": 0.09618223458528519, + "learning_rate": 0.0001235, + "loss": 0.056, + "step": 2470 + }, + { + "epoch": 3.7293233082706765, + "grad_norm": 0.06035883352160454, + "learning_rate": 0.000124, + "loss": 0.0559, + "step": 2480 + }, + { + "epoch": 3.744360902255639, + "grad_norm": 0.06093309447169304, + "learning_rate": 0.0001245, + "loss": 0.0564, + "step": 2490 + }, + { + "epoch": 3.7593984962406015, + "grad_norm": 0.061859678477048874, + "learning_rate": 0.000125, + "loss": 0.0587, + "step": 2500 + }, + { + "epoch": 3.774436090225564, + "grad_norm": 0.08387048542499542, + "learning_rate": 0.00012550000000000001, + "loss": 0.0599, + "step": 2510 + }, + { + "epoch": 3.7894736842105265, + "grad_norm": 0.06999006122350693, + "learning_rate": 0.000126, + "loss": 0.0544, + "step": 2520 + }, + { + "epoch": 3.8045112781954886, + "grad_norm": 0.100027896463871, + "learning_rate": 0.0001265, + "loss": 0.0564, + "step": 2530 + }, + { + "epoch": 3.819548872180451, + "grad_norm": 0.12819582223892212, + "learning_rate": 0.000127, + "loss": 0.0553, + "step": 2540 + }, + { + "epoch": 3.8345864661654137, + "grad_norm": 0.07260995358228683, + "learning_rate": 0.0001275, + "loss": 0.0572, + "step": 2550 + }, + { + "epoch": 3.8496240601503757, + "grad_norm": 0.05892708897590637, + "learning_rate": 0.000128, + "loss": 0.0594, + "step": 2560 + }, + { + "epoch": 3.8646616541353382, + "grad_norm": 0.07078971713781357, + "learning_rate": 0.0001285, + "loss": 0.0556, + "step": 2570 + }, + { + "epoch": 3.8796992481203008, + "grad_norm": 0.06469018757343292, + "learning_rate": 0.00012900000000000002, + "loss": 0.0593, + "step": 2580 + }, + { + "epoch": 3.8947368421052633, + "grad_norm": 0.06514580547809601, + "learning_rate": 0.0001295, + "loss": 0.0537, + "step": 2590 + }, + { + "epoch": 3.909774436090226, + "grad_norm": 0.06876766681671143, + "learning_rate": 0.00013000000000000002, + "loss": 0.0583, + "step": 2600 + }, + { + "epoch": 3.924812030075188, + "grad_norm": 0.08264245837926865, + "learning_rate": 0.0001305, + "loss": 0.0583, + "step": 2610 + }, + { + "epoch": 3.9398496240601504, + "grad_norm": 0.08388067781925201, + "learning_rate": 0.000131, + "loss": 0.0607, + "step": 2620 + }, + { + "epoch": 3.954887218045113, + "grad_norm": 0.06596767157316208, + "learning_rate": 0.0001315, + "loss": 0.0531, + "step": 2630 + }, + { + "epoch": 3.969924812030075, + "grad_norm": 0.07071387767791748, + "learning_rate": 0.000132, + "loss": 0.0594, + "step": 2640 + }, + { + "epoch": 3.9849624060150375, + "grad_norm": 0.08383678644895554, + "learning_rate": 0.00013250000000000002, + "loss": 0.0559, + "step": 2650 + }, + { + "epoch": 4.0, + "grad_norm": 0.0709977000951767, + "learning_rate": 0.000133, + "loss": 0.0555, + "step": 2660 + }, + { + "epoch": 4.015037593984962, + "grad_norm": 0.07539714127779007, + "learning_rate": 0.00013350000000000002, + "loss": 0.0536, + "step": 2670 + }, + { + "epoch": 4.030075187969925, + "grad_norm": 0.06573165208101273, + "learning_rate": 0.000134, + "loss": 0.0534, + "step": 2680 + }, + { + "epoch": 4.045112781954887, + "grad_norm": 0.1408744752407074, + "learning_rate": 0.00013450000000000002, + "loss": 0.0516, + "step": 2690 + }, + { + "epoch": 4.06015037593985, + "grad_norm": 0.11208081990480423, + "learning_rate": 0.000135, + "loss": 0.0521, + "step": 2700 + }, + { + "epoch": 4.075187969924812, + "grad_norm": 0.08213663101196289, + "learning_rate": 0.00013550000000000001, + "loss": 0.055, + "step": 2710 + }, + { + "epoch": 4.090225563909774, + "grad_norm": 0.14577563107013702, + "learning_rate": 0.00013600000000000003, + "loss": 0.0581, + "step": 2720 + }, + { + "epoch": 4.105263157894737, + "grad_norm": 0.10748621076345444, + "learning_rate": 0.0001365, + "loss": 0.0546, + "step": 2730 + }, + { + "epoch": 4.120300751879699, + "grad_norm": 0.0673360526561737, + "learning_rate": 0.00013700000000000002, + "loss": 0.0528, + "step": 2740 + }, + { + "epoch": 4.135338345864661, + "grad_norm": 0.061313945800065994, + "learning_rate": 0.0001375, + "loss": 0.0526, + "step": 2750 + }, + { + "epoch": 4.150375939849624, + "grad_norm": 0.0750400498509407, + "learning_rate": 0.00013800000000000002, + "loss": 0.0493, + "step": 2760 + }, + { + "epoch": 4.165413533834586, + "grad_norm": 0.12614278495311737, + "learning_rate": 0.0001385, + "loss": 0.0595, + "step": 2770 + }, + { + "epoch": 4.180451127819548, + "grad_norm": 0.06822695583105087, + "learning_rate": 0.00013900000000000002, + "loss": 0.0571, + "step": 2780 + }, + { + "epoch": 4.195488721804511, + "grad_norm": 0.07335261255502701, + "learning_rate": 0.0001395, + "loss": 0.0546, + "step": 2790 + }, + { + "epoch": 4.2105263157894735, + "grad_norm": 0.09210097789764404, + "learning_rate": 0.00014000000000000001, + "loss": 0.0553, + "step": 2800 + }, + { + "epoch": 4.225563909774436, + "grad_norm": 0.07518230378627777, + "learning_rate": 0.00014050000000000003, + "loss": 0.0544, + "step": 2810 + }, + { + "epoch": 4.2406015037593985, + "grad_norm": 0.09941935539245605, + "learning_rate": 0.00014099999999999998, + "loss": 0.0495, + "step": 2820 + }, + { + "epoch": 4.2556390977443606, + "grad_norm": 0.07029841840267181, + "learning_rate": 0.0001415, + "loss": 0.0542, + "step": 2830 + }, + { + "epoch": 4.2706766917293235, + "grad_norm": 0.07480385899543762, + "learning_rate": 0.00014199999999999998, + "loss": 0.0566, + "step": 2840 + }, + { + "epoch": 4.285714285714286, + "grad_norm": 0.07338500767946243, + "learning_rate": 0.0001425, + "loss": 0.058, + "step": 2850 + }, + { + "epoch": 4.3007518796992485, + "grad_norm": 0.0869617760181427, + "learning_rate": 0.00014299999999999998, + "loss": 0.0532, + "step": 2860 + }, + { + "epoch": 4.315789473684211, + "grad_norm": 0.06730924546718597, + "learning_rate": 0.0001435, + "loss": 0.0553, + "step": 2870 + }, + { + "epoch": 4.330827067669173, + "grad_norm": 0.06765337288379669, + "learning_rate": 0.000144, + "loss": 0.0521, + "step": 2880 + }, + { + "epoch": 4.345864661654136, + "grad_norm": 0.09918821603059769, + "learning_rate": 0.0001445, + "loss": 0.0462, + "step": 2890 + }, + { + "epoch": 4.360902255639098, + "grad_norm": 0.09746561199426651, + "learning_rate": 0.000145, + "loss": 0.0525, + "step": 2900 + }, + { + "epoch": 4.37593984962406, + "grad_norm": 0.07482302933931351, + "learning_rate": 0.00014549999999999999, + "loss": 0.0474, + "step": 2910 + }, + { + "epoch": 4.390977443609023, + "grad_norm": 0.07530105113983154, + "learning_rate": 0.000146, + "loss": 0.0556, + "step": 2920 + }, + { + "epoch": 4.406015037593985, + "grad_norm": 0.09588643908500671, + "learning_rate": 0.00014649999999999998, + "loss": 0.0546, + "step": 2930 + }, + { + "epoch": 4.421052631578947, + "grad_norm": 0.0772845447063446, + "learning_rate": 0.000147, + "loss": 0.0517, + "step": 2940 + }, + { + "epoch": 4.43609022556391, + "grad_norm": 0.08321461081504822, + "learning_rate": 0.0001475, + "loss": 0.0533, + "step": 2950 + }, + { + "epoch": 4.451127819548872, + "grad_norm": 0.07541036605834961, + "learning_rate": 0.000148, + "loss": 0.051, + "step": 2960 + }, + { + "epoch": 4.466165413533835, + "grad_norm": 0.06388884782791138, + "learning_rate": 0.0001485, + "loss": 0.0521, + "step": 2970 + }, + { + "epoch": 4.481203007518797, + "grad_norm": 0.08425343036651611, + "learning_rate": 0.000149, + "loss": 0.0507, + "step": 2980 + }, + { + "epoch": 4.496240601503759, + "grad_norm": 0.07849389314651489, + "learning_rate": 0.0001495, + "loss": 0.0531, + "step": 2990 + }, + { + "epoch": 4.511278195488722, + "grad_norm": 0.07871508598327637, + "learning_rate": 0.00015, + "loss": 0.056, + "step": 3000 + }, + { + "epoch": 4.511278195488722, + "eval_cer": 0.02158617809991037, + "eval_loss": 0.051458947360515594, + "eval_runtime": 157.7183, + "eval_samples_per_second": 101.618, + "eval_steps_per_second": 0.799, + "eval_wer": 0.07809769024037287, + "step": 3000 + }, + { + "epoch": 4.526315789473684, + "grad_norm": 0.0827491283416748, + "learning_rate": 0.0001505, + "loss": 0.0553, + "step": 3010 + }, + { + "epoch": 4.541353383458647, + "grad_norm": 0.08549568057060242, + "learning_rate": 0.000151, + "loss": 0.0562, + "step": 3020 + }, + { + "epoch": 4.556390977443609, + "grad_norm": 0.08543427288532257, + "learning_rate": 0.0001515, + "loss": 0.0509, + "step": 3030 + }, + { + "epoch": 4.571428571428571, + "grad_norm": 0.06908652931451797, + "learning_rate": 0.000152, + "loss": 0.0474, + "step": 3040 + }, + { + "epoch": 4.586466165413534, + "grad_norm": 0.08035128563642502, + "learning_rate": 0.0001525, + "loss": 0.0486, + "step": 3050 + }, + { + "epoch": 4.601503759398496, + "grad_norm": 0.06270065158605576, + "learning_rate": 0.000153, + "loss": 0.0499, + "step": 3060 + }, + { + "epoch": 4.616541353383458, + "grad_norm": 0.12812690436840057, + "learning_rate": 0.0001535, + "loss": 0.0559, + "step": 3070 + }, + { + "epoch": 4.631578947368421, + "grad_norm": 0.05992261692881584, + "learning_rate": 0.000154, + "loss": 0.0526, + "step": 3080 + }, + { + "epoch": 4.646616541353383, + "grad_norm": 0.07544948905706406, + "learning_rate": 0.00015450000000000001, + "loss": 0.0487, + "step": 3090 + }, + { + "epoch": 4.661654135338345, + "grad_norm": 0.07155109941959381, + "learning_rate": 0.000155, + "loss": 0.0487, + "step": 3100 + }, + { + "epoch": 4.676691729323308, + "grad_norm": 0.07207630574703217, + "learning_rate": 0.0001555, + "loss": 0.051, + "step": 3110 + }, + { + "epoch": 4.69172932330827, + "grad_norm": 0.05290692672133446, + "learning_rate": 0.000156, + "loss": 0.0532, + "step": 3120 + }, + { + "epoch": 4.706766917293233, + "grad_norm": 0.0729108527302742, + "learning_rate": 0.0001565, + "loss": 0.0545, + "step": 3130 + }, + { + "epoch": 4.7218045112781954, + "grad_norm": 0.08135359734296799, + "learning_rate": 0.000157, + "loss": 0.0499, + "step": 3140 + }, + { + "epoch": 4.7368421052631575, + "grad_norm": 0.0715440958738327, + "learning_rate": 0.0001575, + "loss": 0.05, + "step": 3150 + }, + { + "epoch": 4.7518796992481205, + "grad_norm": 0.09423758089542389, + "learning_rate": 0.000158, + "loss": 0.0538, + "step": 3160 + }, + { + "epoch": 4.7669172932330826, + "grad_norm": 0.059924185276031494, + "learning_rate": 0.0001585, + "loss": 0.0481, + "step": 3170 + }, + { + "epoch": 4.7819548872180455, + "grad_norm": 0.10351450741291046, + "learning_rate": 0.00015900000000000002, + "loss": 0.045, + "step": 3180 + }, + { + "epoch": 4.796992481203008, + "grad_norm": 0.061447639018297195, + "learning_rate": 0.0001595, + "loss": 0.0489, + "step": 3190 + }, + { + "epoch": 4.81203007518797, + "grad_norm": 0.06320513784885406, + "learning_rate": 0.00016, + "loss": 0.0472, + "step": 3200 + }, + { + "epoch": 4.827067669172933, + "grad_norm": 0.07715082168579102, + "learning_rate": 0.0001605, + "loss": 0.0495, + "step": 3210 + }, + { + "epoch": 4.842105263157895, + "grad_norm": 0.1049925908446312, + "learning_rate": 0.000161, + "loss": 0.0477, + "step": 3220 + }, + { + "epoch": 4.857142857142857, + "grad_norm": 0.0546201728284359, + "learning_rate": 0.0001615, + "loss": 0.0464, + "step": 3230 + }, + { + "epoch": 4.87218045112782, + "grad_norm": 0.06560751795768738, + "learning_rate": 0.000162, + "loss": 0.0482, + "step": 3240 + }, + { + "epoch": 4.887218045112782, + "grad_norm": 0.06029576063156128, + "learning_rate": 0.00016250000000000002, + "loss": 0.0487, + "step": 3250 + }, + { + "epoch": 4.902255639097744, + "grad_norm": 0.06343934684991837, + "learning_rate": 0.000163, + "loss": 0.0486, + "step": 3260 + }, + { + "epoch": 4.917293233082707, + "grad_norm": 0.08357163518667221, + "learning_rate": 0.00016350000000000002, + "loss": 0.047, + "step": 3270 + }, + { + "epoch": 4.932330827067669, + "grad_norm": 0.06959404051303864, + "learning_rate": 0.000164, + "loss": 0.0449, + "step": 3280 + }, + { + "epoch": 4.947368421052632, + "grad_norm": 0.06669142097234726, + "learning_rate": 0.00016450000000000001, + "loss": 0.0481, + "step": 3290 + }, + { + "epoch": 4.962406015037594, + "grad_norm": 0.06286276876926422, + "learning_rate": 0.000165, + "loss": 0.0486, + "step": 3300 + }, + { + "epoch": 4.977443609022556, + "grad_norm": 0.07344794273376465, + "learning_rate": 0.0001655, + "loss": 0.0474, + "step": 3310 + }, + { + "epoch": 4.992481203007519, + "grad_norm": 0.055243995040655136, + "learning_rate": 0.00016600000000000002, + "loss": 0.047, + "step": 3320 + }, + { + "epoch": 5.007518796992481, + "grad_norm": 0.06104254722595215, + "learning_rate": 0.0001665, + "loss": 0.0476, + "step": 3330 + }, + { + "epoch": 5.022556390977444, + "grad_norm": 0.07689296454191208, + "learning_rate": 0.00016700000000000002, + "loss": 0.0452, + "step": 3340 + }, + { + "epoch": 5.037593984962406, + "grad_norm": 0.07180420309305191, + "learning_rate": 0.0001675, + "loss": 0.049, + "step": 3350 + }, + { + "epoch": 5.052631578947368, + "grad_norm": 0.06476438790559769, + "learning_rate": 0.00016800000000000002, + "loss": 0.0413, + "step": 3360 + }, + { + "epoch": 5.067669172932331, + "grad_norm": 0.06656920164823532, + "learning_rate": 0.0001685, + "loss": 0.0445, + "step": 3370 + }, + { + "epoch": 5.082706766917293, + "grad_norm": 0.0829206332564354, + "learning_rate": 0.00016900000000000002, + "loss": 0.048, + "step": 3380 + }, + { + "epoch": 5.097744360902255, + "grad_norm": 0.07200006395578384, + "learning_rate": 0.00016950000000000003, + "loss": 0.046, + "step": 3390 + }, + { + "epoch": 5.112781954887218, + "grad_norm": 0.05968916788697243, + "learning_rate": 0.00017, + "loss": 0.0471, + "step": 3400 + }, + { + "epoch": 5.12781954887218, + "grad_norm": 0.08708423376083374, + "learning_rate": 0.00017050000000000002, + "loss": 0.0447, + "step": 3410 + }, + { + "epoch": 5.142857142857143, + "grad_norm": 0.0959819108247757, + "learning_rate": 0.000171, + "loss": 0.0436, + "step": 3420 + }, + { + "epoch": 5.157894736842105, + "grad_norm": 0.05492165684700012, + "learning_rate": 0.00017150000000000002, + "loss": 0.0453, + "step": 3430 + }, + { + "epoch": 5.172932330827067, + "grad_norm": 0.061032190918922424, + "learning_rate": 0.00017199999999999998, + "loss": 0.0492, + "step": 3440 + }, + { + "epoch": 5.18796992481203, + "grad_norm": 0.0586761049926281, + "learning_rate": 0.0001725, + "loss": 0.0444, + "step": 3450 + }, + { + "epoch": 5.203007518796992, + "grad_norm": 0.06914158910512924, + "learning_rate": 0.000173, + "loss": 0.0457, + "step": 3460 + }, + { + "epoch": 5.2180451127819545, + "grad_norm": 0.06396849453449249, + "learning_rate": 0.0001735, + "loss": 0.0503, + "step": 3470 + }, + { + "epoch": 5.2330827067669174, + "grad_norm": 0.06256670504808426, + "learning_rate": 0.000174, + "loss": 0.0459, + "step": 3480 + }, + { + "epoch": 5.2481203007518795, + "grad_norm": 0.06614803522825241, + "learning_rate": 0.00017449999999999999, + "loss": 0.0419, + "step": 3490 + }, + { + "epoch": 5.2631578947368425, + "grad_norm": 0.10315285623073578, + "learning_rate": 0.000175, + "loss": 0.0497, + "step": 3500 + }, + { + "epoch": 5.2781954887218046, + "grad_norm": 0.06165418401360512, + "learning_rate": 0.00017549999999999998, + "loss": 0.0426, + "step": 3510 + }, + { + "epoch": 5.293233082706767, + "grad_norm": 0.059949927031993866, + "learning_rate": 0.000176, + "loss": 0.0442, + "step": 3520 + }, + { + "epoch": 5.30827067669173, + "grad_norm": 0.08298108726739883, + "learning_rate": 0.00017649999999999998, + "loss": 0.0457, + "step": 3530 + }, + { + "epoch": 5.323308270676692, + "grad_norm": 0.07161926478147507, + "learning_rate": 0.000177, + "loss": 0.0479, + "step": 3540 + }, + { + "epoch": 5.338345864661654, + "grad_norm": 0.06794701516628265, + "learning_rate": 0.0001775, + "loss": 0.045, + "step": 3550 + }, + { + "epoch": 5.353383458646617, + "grad_norm": 0.06527054309844971, + "learning_rate": 0.000178, + "loss": 0.0444, + "step": 3560 + }, + { + "epoch": 5.368421052631579, + "grad_norm": 0.06878358870744705, + "learning_rate": 0.0001785, + "loss": 0.0434, + "step": 3570 + }, + { + "epoch": 5.383458646616542, + "grad_norm": 0.057137925177812576, + "learning_rate": 0.000179, + "loss": 0.0446, + "step": 3580 + }, + { + "epoch": 5.398496240601504, + "grad_norm": 0.07706617563962936, + "learning_rate": 0.0001795, + "loss": 0.0436, + "step": 3590 + }, + { + "epoch": 5.413533834586466, + "grad_norm": 0.09341507405042648, + "learning_rate": 0.00017999999999999998, + "loss": 0.0431, + "step": 3600 + }, + { + "epoch": 5.428571428571429, + "grad_norm": 0.12199874222278595, + "learning_rate": 0.0001805, + "loss": 0.0443, + "step": 3610 + }, + { + "epoch": 5.443609022556391, + "grad_norm": 0.06663978844881058, + "learning_rate": 0.000181, + "loss": 0.0466, + "step": 3620 + }, + { + "epoch": 5.458646616541353, + "grad_norm": 0.11655236780643463, + "learning_rate": 0.0001815, + "loss": 0.0434, + "step": 3630 + }, + { + "epoch": 5.473684210526316, + "grad_norm": 0.05978340655565262, + "learning_rate": 0.000182, + "loss": 0.0443, + "step": 3640 + }, + { + "epoch": 5.488721804511278, + "grad_norm": 0.05883520469069481, + "learning_rate": 0.0001825, + "loss": 0.0442, + "step": 3650 + }, + { + "epoch": 5.503759398496241, + "grad_norm": 0.0803222507238388, + "learning_rate": 0.000183, + "loss": 0.0402, + "step": 3660 + }, + { + "epoch": 5.518796992481203, + "grad_norm": 0.06758495420217514, + "learning_rate": 0.0001835, + "loss": 0.0446, + "step": 3670 + }, + { + "epoch": 5.533834586466165, + "grad_norm": 0.07356420159339905, + "learning_rate": 0.000184, + "loss": 0.0434, + "step": 3680 + }, + { + "epoch": 5.548872180451128, + "grad_norm": 0.06145314499735832, + "learning_rate": 0.0001845, + "loss": 0.0441, + "step": 3690 + }, + { + "epoch": 5.56390977443609, + "grad_norm": 0.05586985871195793, + "learning_rate": 0.000185, + "loss": 0.039, + "step": 3700 + }, + { + "epoch": 5.578947368421053, + "grad_norm": 0.1432076394557953, + "learning_rate": 0.0001855, + "loss": 0.0486, + "step": 3710 + }, + { + "epoch": 5.593984962406015, + "grad_norm": 0.06590097397565842, + "learning_rate": 0.000186, + "loss": 0.0487, + "step": 3720 + }, + { + "epoch": 5.609022556390977, + "grad_norm": 0.06352625042200089, + "learning_rate": 0.0001865, + "loss": 0.0443, + "step": 3730 + }, + { + "epoch": 5.62406015037594, + "grad_norm": 0.07959103584289551, + "learning_rate": 0.000187, + "loss": 0.0452, + "step": 3740 + }, + { + "epoch": 5.639097744360902, + "grad_norm": 0.06270395964384079, + "learning_rate": 0.0001875, + "loss": 0.0399, + "step": 3750 + }, + { + "epoch": 5.654135338345864, + "grad_norm": 0.06196174398064613, + "learning_rate": 0.00018800000000000002, + "loss": 0.04, + "step": 3760 + }, + { + "epoch": 5.669172932330827, + "grad_norm": 0.06072288006544113, + "learning_rate": 0.0001885, + "loss": 0.0454, + "step": 3770 + }, + { + "epoch": 5.684210526315789, + "grad_norm": 0.05780434608459473, + "learning_rate": 0.000189, + "loss": 0.0457, + "step": 3780 + }, + { + "epoch": 5.6992481203007515, + "grad_norm": 0.056580331176519394, + "learning_rate": 0.0001895, + "loss": 0.0439, + "step": 3790 + }, + { + "epoch": 5.714285714285714, + "grad_norm": 0.06780938059091568, + "learning_rate": 0.00019, + "loss": 0.042, + "step": 3800 + }, + { + "epoch": 5.7293233082706765, + "grad_norm": 0.07645943015813828, + "learning_rate": 0.0001905, + "loss": 0.0436, + "step": 3810 + }, + { + "epoch": 5.7443609022556394, + "grad_norm": 0.06665679067373276, + "learning_rate": 0.000191, + "loss": 0.046, + "step": 3820 + }, + { + "epoch": 5.7593984962406015, + "grad_norm": 0.06348271667957306, + "learning_rate": 0.00019150000000000002, + "loss": 0.0436, + "step": 3830 + }, + { + "epoch": 5.774436090225564, + "grad_norm": 0.07240138947963715, + "learning_rate": 0.000192, + "loss": 0.0434, + "step": 3840 + }, + { + "epoch": 5.7894736842105265, + "grad_norm": 0.07741253077983856, + "learning_rate": 0.00019250000000000002, + "loss": 0.0408, + "step": 3850 + }, + { + "epoch": 5.804511278195489, + "grad_norm": 0.06330270320177078, + "learning_rate": 0.000193, + "loss": 0.0391, + "step": 3860 + }, + { + "epoch": 5.819548872180452, + "grad_norm": 0.08810721337795258, + "learning_rate": 0.00019350000000000001, + "loss": 0.0419, + "step": 3870 + }, + { + "epoch": 5.834586466165414, + "grad_norm": 0.08770038187503815, + "learning_rate": 0.000194, + "loss": 0.042, + "step": 3880 + }, + { + "epoch": 5.849624060150376, + "grad_norm": 0.09694177657365799, + "learning_rate": 0.0001945, + "loss": 0.0483, + "step": 3890 + }, + { + "epoch": 5.864661654135339, + "grad_norm": 0.06397241353988647, + "learning_rate": 0.00019500000000000002, + "loss": 0.0443, + "step": 3900 + }, + { + "epoch": 5.879699248120301, + "grad_norm": 0.06331652402877808, + "learning_rate": 0.0001955, + "loss": 0.0454, + "step": 3910 + }, + { + "epoch": 5.894736842105263, + "grad_norm": 0.05943892523646355, + "learning_rate": 0.00019600000000000002, + "loss": 0.042, + "step": 3920 + }, + { + "epoch": 5.909774436090226, + "grad_norm": 0.07324781268835068, + "learning_rate": 0.0001965, + "loss": 0.0435, + "step": 3930 + }, + { + "epoch": 5.924812030075188, + "grad_norm": 0.06800910830497742, + "learning_rate": 0.00019700000000000002, + "loss": 0.0402, + "step": 3940 + }, + { + "epoch": 5.93984962406015, + "grad_norm": 0.06812734156847, + "learning_rate": 0.0001975, + "loss": 0.0421, + "step": 3950 + }, + { + "epoch": 5.954887218045113, + "grad_norm": 0.06211187690496445, + "learning_rate": 0.00019800000000000002, + "loss": 0.0497, + "step": 3960 + }, + { + "epoch": 5.969924812030075, + "grad_norm": 0.08975367248058319, + "learning_rate": 0.00019850000000000003, + "loss": 0.0444, + "step": 3970 + }, + { + "epoch": 5.984962406015038, + "grad_norm": 0.076654814183712, + "learning_rate": 0.000199, + "loss": 0.0459, + "step": 3980 + }, + { + "epoch": 6.0, + "grad_norm": 0.07650908082723618, + "learning_rate": 0.00019950000000000002, + "loss": 0.0475, + "step": 3990 + }, + { + "epoch": 6.015037593984962, + "grad_norm": 0.06910504400730133, + "learning_rate": 0.0002, + "loss": 0.0423, + "step": 4000 + }, + { + "epoch": 6.015037593984962, + "eval_cer": 0.020727273716437785, + "eval_loss": 0.04866219311952591, + "eval_runtime": 161.7042, + "eval_samples_per_second": 99.113, + "eval_steps_per_second": 0.779, + "eval_wer": 0.07262869063504294, + "step": 4000 + }, + { + "epoch": 6.030075187969925, + "grad_norm": 0.05724914371967316, + "learning_rate": 0.00020050000000000002, + "loss": 0.0402, + "step": 4010 + }, + { + "epoch": 6.045112781954887, + "grad_norm": 0.06336195766925812, + "learning_rate": 0.000201, + "loss": 0.0421, + "step": 4020 + }, + { + "epoch": 6.06015037593985, + "grad_norm": 0.07029417902231216, + "learning_rate": 0.00020150000000000002, + "loss": 0.0426, + "step": 4030 + }, + { + "epoch": 6.075187969924812, + "grad_norm": 0.05938456580042839, + "learning_rate": 0.000202, + "loss": 0.0381, + "step": 4040 + }, + { + "epoch": 6.090225563909774, + "grad_norm": 0.06267672777175903, + "learning_rate": 0.00020250000000000002, + "loss": 0.0412, + "step": 4050 + }, + { + "epoch": 6.105263157894737, + "grad_norm": 0.0631110891699791, + "learning_rate": 0.00020300000000000003, + "loss": 0.0389, + "step": 4060 + }, + { + "epoch": 6.120300751879699, + "grad_norm": 0.0681823343038559, + "learning_rate": 0.00020349999999999999, + "loss": 0.037, + "step": 4070 + }, + { + "epoch": 6.135338345864661, + "grad_norm": 0.05891154706478119, + "learning_rate": 0.000204, + "loss": 0.042, + "step": 4080 + }, + { + "epoch": 6.150375939849624, + "grad_norm": 0.06000526249408722, + "learning_rate": 0.00020449999999999998, + "loss": 0.0419, + "step": 4090 + }, + { + "epoch": 6.165413533834586, + "grad_norm": 0.06349631398916245, + "learning_rate": 0.000205, + "loss": 0.0376, + "step": 4100 + }, + { + "epoch": 6.180451127819548, + "grad_norm": 0.05860697478055954, + "learning_rate": 0.00020549999999999998, + "loss": 0.0396, + "step": 4110 + }, + { + "epoch": 6.195488721804511, + "grad_norm": 0.056414201855659485, + "learning_rate": 0.000206, + "loss": 0.0446, + "step": 4120 + }, + { + "epoch": 6.2105263157894735, + "grad_norm": 0.0755389928817749, + "learning_rate": 0.0002065, + "loss": 0.0406, + "step": 4130 + }, + { + "epoch": 6.225563909774436, + "grad_norm": 0.06304564327001572, + "learning_rate": 0.000207, + "loss": 0.0386, + "step": 4140 + }, + { + "epoch": 6.2406015037593985, + "grad_norm": 0.05802801251411438, + "learning_rate": 0.0002075, + "loss": 0.0404, + "step": 4150 + }, + { + "epoch": 6.2556390977443606, + "grad_norm": 0.07250794768333435, + "learning_rate": 0.000208, + "loss": 0.041, + "step": 4160 + }, + { + "epoch": 6.2706766917293235, + "grad_norm": 0.10487314313650131, + "learning_rate": 0.0002085, + "loss": 0.04, + "step": 4170 + }, + { + "epoch": 6.285714285714286, + "grad_norm": 0.11122310161590576, + "learning_rate": 0.00020899999999999998, + "loss": 0.0393, + "step": 4180 + }, + { + "epoch": 6.3007518796992485, + "grad_norm": 0.09853117167949677, + "learning_rate": 0.0002095, + "loss": 0.0431, + "step": 4190 + }, + { + "epoch": 6.315789473684211, + "grad_norm": 0.0656944066286087, + "learning_rate": 0.00021, + "loss": 0.0388, + "step": 4200 + }, + { + "epoch": 6.330827067669173, + "grad_norm": 0.06939065456390381, + "learning_rate": 0.0002105, + "loss": 0.0387, + "step": 4210 + }, + { + "epoch": 6.345864661654136, + "grad_norm": 0.0631631538271904, + "learning_rate": 0.000211, + "loss": 0.0404, + "step": 4220 + }, + { + "epoch": 6.360902255639098, + "grad_norm": 0.08743909746408463, + "learning_rate": 0.0002115, + "loss": 0.0407, + "step": 4230 + }, + { + "epoch": 6.37593984962406, + "grad_norm": 0.0623641200363636, + "learning_rate": 0.000212, + "loss": 0.0404, + "step": 4240 + }, + { + "epoch": 6.390977443609023, + "grad_norm": 0.0821632593870163, + "learning_rate": 0.0002125, + "loss": 0.0377, + "step": 4250 + }, + { + "epoch": 6.406015037593985, + "grad_norm": 0.06677273660898209, + "learning_rate": 0.000213, + "loss": 0.0344, + "step": 4260 + }, + { + "epoch": 6.421052631578947, + "grad_norm": 0.07842960953712463, + "learning_rate": 0.0002135, + "loss": 0.0402, + "step": 4270 + }, + { + "epoch": 6.43609022556391, + "grad_norm": 0.061761919409036636, + "learning_rate": 0.000214, + "loss": 0.0353, + "step": 4280 + }, + { + "epoch": 6.451127819548872, + "grad_norm": 0.07739520817995071, + "learning_rate": 0.0002145, + "loss": 0.0395, + "step": 4290 + }, + { + "epoch": 6.466165413533835, + "grad_norm": 0.07733475416898727, + "learning_rate": 0.000215, + "loss": 0.0361, + "step": 4300 + }, + { + "epoch": 6.481203007518797, + "grad_norm": 0.0634097158908844, + "learning_rate": 0.0002155, + "loss": 0.0379, + "step": 4310 + }, + { + "epoch": 6.496240601503759, + "grad_norm": 0.07296420633792877, + "learning_rate": 0.000216, + "loss": 0.0414, + "step": 4320 + }, + { + "epoch": 6.511278195488722, + "grad_norm": 0.06225402280688286, + "learning_rate": 0.0002165, + "loss": 0.0336, + "step": 4330 + }, + { + "epoch": 6.526315789473684, + "grad_norm": 0.06256049126386642, + "learning_rate": 0.00021700000000000002, + "loss": 0.04, + "step": 4340 + }, + { + "epoch": 6.541353383458647, + "grad_norm": 0.057943835854530334, + "learning_rate": 0.0002175, + "loss": 0.0392, + "step": 4350 + }, + { + "epoch": 6.556390977443609, + "grad_norm": 0.06446385383605957, + "learning_rate": 0.000218, + "loss": 0.0372, + "step": 4360 + }, + { + "epoch": 6.571428571428571, + "grad_norm": 0.08455291390419006, + "learning_rate": 0.0002185, + "loss": 0.0405, + "step": 4370 + }, + { + "epoch": 6.586466165413534, + "grad_norm": 0.05729060620069504, + "learning_rate": 0.000219, + "loss": 0.0389, + "step": 4380 + }, + { + "epoch": 6.601503759398496, + "grad_norm": 0.1113635003566742, + "learning_rate": 0.0002195, + "loss": 0.0385, + "step": 4390 + }, + { + "epoch": 6.616541353383458, + "grad_norm": 0.07287126034498215, + "learning_rate": 0.00022, + "loss": 0.038, + "step": 4400 + }, + { + "epoch": 6.631578947368421, + "grad_norm": 0.05508929118514061, + "learning_rate": 0.0002205, + "loss": 0.0356, + "step": 4410 + }, + { + "epoch": 6.646616541353383, + "grad_norm": 0.0543157234787941, + "learning_rate": 0.000221, + "loss": 0.0385, + "step": 4420 + }, + { + "epoch": 6.661654135338345, + "grad_norm": 0.05844772979617119, + "learning_rate": 0.00022150000000000002, + "loss": 0.0381, + "step": 4430 + }, + { + "epoch": 6.676691729323308, + "grad_norm": 0.06718626618385315, + "learning_rate": 0.000222, + "loss": 0.0359, + "step": 4440 + }, + { + "epoch": 6.69172932330827, + "grad_norm": 0.05742814391851425, + "learning_rate": 0.00022250000000000001, + "loss": 0.0373, + "step": 4450 + }, + { + "epoch": 6.706766917293233, + "grad_norm": 0.08460763096809387, + "learning_rate": 0.000223, + "loss": 0.034, + "step": 4460 + }, + { + "epoch": 6.7218045112781954, + "grad_norm": 0.06337951123714447, + "learning_rate": 0.0002235, + "loss": 0.039, + "step": 4470 + }, + { + "epoch": 6.7368421052631575, + "grad_norm": 0.0621739998459816, + "learning_rate": 0.000224, + "loss": 0.0362, + "step": 4480 + }, + { + "epoch": 6.7518796992481205, + "grad_norm": 0.08055991679430008, + "learning_rate": 0.0002245, + "loss": 0.0384, + "step": 4490 + }, + { + "epoch": 6.7669172932330826, + "grad_norm": 0.04713355377316475, + "learning_rate": 0.00022500000000000002, + "loss": 0.037, + "step": 4500 + }, + { + "epoch": 6.7819548872180455, + "grad_norm": 0.06481621414422989, + "learning_rate": 0.0002255, + "loss": 0.0408, + "step": 4510 + }, + { + "epoch": 6.796992481203008, + "grad_norm": 0.08068499714136124, + "learning_rate": 0.00022600000000000002, + "loss": 0.0406, + "step": 4520 + }, + { + "epoch": 6.81203007518797, + "grad_norm": 0.06626643240451813, + "learning_rate": 0.0002265, + "loss": 0.0381, + "step": 4530 + }, + { + "epoch": 6.827067669172933, + "grad_norm": 0.06152673810720444, + "learning_rate": 0.00022700000000000002, + "loss": 0.0381, + "step": 4540 + }, + { + "epoch": 6.842105263157895, + "grad_norm": 0.10953982919454575, + "learning_rate": 0.0002275, + "loss": 0.0384, + "step": 4550 + }, + { + "epoch": 6.857142857142857, + "grad_norm": 0.08171266317367554, + "learning_rate": 0.000228, + "loss": 0.0335, + "step": 4560 + }, + { + "epoch": 6.87218045112782, + "grad_norm": 0.05440454185009003, + "learning_rate": 0.00022850000000000002, + "loss": 0.0376, + "step": 4570 + }, + { + "epoch": 6.887218045112782, + "grad_norm": 0.0660145953297615, + "learning_rate": 0.000229, + "loss": 0.0371, + "step": 4580 + }, + { + "epoch": 6.902255639097744, + "grad_norm": 0.07334434241056442, + "learning_rate": 0.00022950000000000002, + "loss": 0.0412, + "step": 4590 + }, + { + "epoch": 6.917293233082707, + "grad_norm": 0.04670143872499466, + "learning_rate": 0.00023, + "loss": 0.0351, + "step": 4600 + }, + { + "epoch": 6.932330827067669, + "grad_norm": 0.0678081139922142, + "learning_rate": 0.00023050000000000002, + "loss": 0.0413, + "step": 4610 + }, + { + "epoch": 6.947368421052632, + "grad_norm": 0.05912181735038757, + "learning_rate": 0.000231, + "loss": 0.0363, + "step": 4620 + }, + { + "epoch": 6.962406015037594, + "grad_norm": 0.06157950311899185, + "learning_rate": 0.00023150000000000002, + "loss": 0.0383, + "step": 4630 + }, + { + "epoch": 6.977443609022556, + "grad_norm": 0.06027334928512573, + "learning_rate": 0.00023200000000000003, + "loss": 0.0403, + "step": 4640 + }, + { + "epoch": 6.992481203007519, + "grad_norm": 0.06633207201957703, + "learning_rate": 0.0002325, + "loss": 0.033, + "step": 4650 + }, + { + "epoch": 7.007518796992481, + "grad_norm": 0.05143498256802559, + "learning_rate": 0.00023300000000000003, + "loss": 0.0359, + "step": 4660 + }, + { + "epoch": 7.022556390977444, + "grad_norm": 0.06413625925779343, + "learning_rate": 0.0002335, + "loss": 0.0343, + "step": 4670 + }, + { + "epoch": 7.037593984962406, + "grad_norm": 0.05680365487933159, + "learning_rate": 0.00023400000000000002, + "loss": 0.0338, + "step": 4680 + }, + { + "epoch": 7.052631578947368, + "grad_norm": 0.06461241841316223, + "learning_rate": 0.00023449999999999998, + "loss": 0.0356, + "step": 4690 + }, + { + "epoch": 7.067669172932331, + "grad_norm": 0.05739262327551842, + "learning_rate": 0.000235, + "loss": 0.0348, + "step": 4700 + }, + { + "epoch": 7.082706766917293, + "grad_norm": 0.061877116560935974, + "learning_rate": 0.0002355, + "loss": 0.0389, + "step": 4710 + }, + { + "epoch": 7.097744360902255, + "grad_norm": 0.07602377980947495, + "learning_rate": 0.000236, + "loss": 0.0378, + "step": 4720 + }, + { + "epoch": 7.112781954887218, + "grad_norm": 0.05630328506231308, + "learning_rate": 0.0002365, + "loss": 0.0339, + "step": 4730 + }, + { + "epoch": 7.12781954887218, + "grad_norm": 0.07818866521120071, + "learning_rate": 0.000237, + "loss": 0.032, + "step": 4740 + }, + { + "epoch": 7.142857142857143, + "grad_norm": 0.0829167515039444, + "learning_rate": 0.0002375, + "loss": 0.0314, + "step": 4750 + }, + { + "epoch": 7.157894736842105, + "grad_norm": 0.06386187672615051, + "learning_rate": 0.00023799999999999998, + "loss": 0.0366, + "step": 4760 + }, + { + "epoch": 7.172932330827067, + "grad_norm": 0.077994205057621, + "learning_rate": 0.0002385, + "loss": 0.0318, + "step": 4770 + }, + { + "epoch": 7.18796992481203, + "grad_norm": 0.05389230698347092, + "learning_rate": 0.00023899999999999998, + "loss": 0.0345, + "step": 4780 + }, + { + "epoch": 7.203007518796992, + "grad_norm": 0.0837327316403389, + "learning_rate": 0.0002395, + "loss": 0.032, + "step": 4790 + }, + { + "epoch": 7.2180451127819545, + "grad_norm": 0.04763004183769226, + "learning_rate": 0.00024, + "loss": 0.0295, + "step": 4800 + }, + { + "epoch": 7.2330827067669174, + "grad_norm": 0.07888182997703552, + "learning_rate": 0.0002405, + "loss": 0.0348, + "step": 4810 + }, + { + "epoch": 7.2481203007518795, + "grad_norm": 0.13571782410144806, + "learning_rate": 0.000241, + "loss": 0.033, + "step": 4820 + }, + { + "epoch": 7.2631578947368425, + "grad_norm": 0.05507836490869522, + "learning_rate": 0.0002415, + "loss": 0.0355, + "step": 4830 + }, + { + "epoch": 7.2781954887218046, + "grad_norm": 0.07828940451145172, + "learning_rate": 0.000242, + "loss": 0.0356, + "step": 4840 + }, + { + "epoch": 7.293233082706767, + "grad_norm": 0.0593411959707737, + "learning_rate": 0.00024249999999999999, + "loss": 0.0319, + "step": 4850 + }, + { + "epoch": 7.30827067669173, + "grad_norm": 0.059401631355285645, + "learning_rate": 0.000243, + "loss": 0.0358, + "step": 4860 + }, + { + "epoch": 7.323308270676692, + "grad_norm": 0.07558241486549377, + "learning_rate": 0.0002435, + "loss": 0.0327, + "step": 4870 + }, + { + "epoch": 7.338345864661654, + "grad_norm": 0.06347363442182541, + "learning_rate": 0.000244, + "loss": 0.0343, + "step": 4880 + }, + { + "epoch": 7.353383458646617, + "grad_norm": 0.08498407900333405, + "learning_rate": 0.0002445, + "loss": 0.0327, + "step": 4890 + }, + { + "epoch": 7.368421052631579, + "grad_norm": 0.06663122773170471, + "learning_rate": 0.000245, + "loss": 0.0341, + "step": 4900 + }, + { + "epoch": 7.383458646616542, + "grad_norm": 0.05973844230175018, + "learning_rate": 0.0002455, + "loss": 0.0347, + "step": 4910 + }, + { + "epoch": 7.398496240601504, + "grad_norm": 0.06749178469181061, + "learning_rate": 0.000246, + "loss": 0.0347, + "step": 4920 + }, + { + "epoch": 7.413533834586466, + "grad_norm": 0.060836948454380035, + "learning_rate": 0.00024650000000000003, + "loss": 0.0328, + "step": 4930 + }, + { + "epoch": 7.428571428571429, + "grad_norm": 0.06503672152757645, + "learning_rate": 0.000247, + "loss": 0.0318, + "step": 4940 + }, + { + "epoch": 7.443609022556391, + "grad_norm": 0.07955773174762726, + "learning_rate": 0.0002475, + "loss": 0.0334, + "step": 4950 + }, + { + "epoch": 7.458646616541353, + "grad_norm": 0.06897212564945221, + "learning_rate": 0.000248, + "loss": 0.0374, + "step": 4960 + }, + { + "epoch": 7.473684210526316, + "grad_norm": 0.04223218187689781, + "learning_rate": 0.0002485, + "loss": 0.03, + "step": 4970 + }, + { + "epoch": 7.488721804511278, + "grad_norm": 0.09226986020803452, + "learning_rate": 0.000249, + "loss": 0.0325, + "step": 4980 + }, + { + "epoch": 7.503759398496241, + "grad_norm": 0.07418844848871231, + "learning_rate": 0.0002495, + "loss": 0.0313, + "step": 4990 + }, + { + "epoch": 7.518796992481203, + "grad_norm": 0.06457831710577011, + "learning_rate": 0.00025, + "loss": 0.0345, + "step": 5000 + }, + { + "epoch": 7.518796992481203, + "eval_cer": 0.019920053207188412, + "eval_loss": 0.04903838038444519, + "eval_runtime": 161.2562, + "eval_samples_per_second": 99.388, + "eval_steps_per_second": 0.781, + "eval_wer": 0.06973914187449491, + "step": 5000 + }, + { + "epoch": 7.533834586466165, + "grad_norm": 0.06083071976900101, + "learning_rate": 0.0002505, + "loss": 0.0362, + "step": 5010 + }, + { + "epoch": 7.548872180451128, + "grad_norm": 0.06106117367744446, + "learning_rate": 0.00025100000000000003, + "loss": 0.0336, + "step": 5020 + }, + { + "epoch": 7.56390977443609, + "grad_norm": 0.059997886419296265, + "learning_rate": 0.0002515, + "loss": 0.0345, + "step": 5030 + }, + { + "epoch": 7.578947368421053, + "grad_norm": 0.271132230758667, + "learning_rate": 0.000252, + "loss": 0.0335, + "step": 5040 + }, + { + "epoch": 7.593984962406015, + "grad_norm": 0.05776393413543701, + "learning_rate": 0.0002525, + "loss": 0.0328, + "step": 5050 + }, + { + "epoch": 7.609022556390977, + "grad_norm": 0.08932015299797058, + "learning_rate": 0.000253, + "loss": 0.0334, + "step": 5060 + }, + { + "epoch": 7.62406015037594, + "grad_norm": 0.07735109329223633, + "learning_rate": 0.0002535, + "loss": 0.0362, + "step": 5070 + }, + { + "epoch": 7.639097744360902, + "grad_norm": 0.06020679697394371, + "learning_rate": 0.000254, + "loss": 0.0349, + "step": 5080 + }, + { + "epoch": 7.654135338345864, + "grad_norm": 0.05556226894259453, + "learning_rate": 0.0002545, + "loss": 0.0363, + "step": 5090 + }, + { + "epoch": 7.669172932330827, + "grad_norm": 0.07329504936933517, + "learning_rate": 0.000255, + "loss": 0.0339, + "step": 5100 + }, + { + "epoch": 7.684210526315789, + "grad_norm": 0.061310190707445145, + "learning_rate": 0.00025550000000000003, + "loss": 0.0343, + "step": 5110 + }, + { + "epoch": 7.6992481203007515, + "grad_norm": 0.08992335945367813, + "learning_rate": 0.000256, + "loss": 0.0337, + "step": 5120 + }, + { + "epoch": 7.714285714285714, + "grad_norm": 0.04959482327103615, + "learning_rate": 0.0002565, + "loss": 0.0356, + "step": 5130 + }, + { + "epoch": 7.7293233082706765, + "grad_norm": 0.07245901972055435, + "learning_rate": 0.000257, + "loss": 0.0302, + "step": 5140 + }, + { + "epoch": 7.7443609022556394, + "grad_norm": 0.07242570072412491, + "learning_rate": 0.0002575, + "loss": 0.0319, + "step": 5150 + }, + { + "epoch": 7.7593984962406015, + "grad_norm": 0.05172485113143921, + "learning_rate": 0.00025800000000000004, + "loss": 0.0367, + "step": 5160 + }, + { + "epoch": 7.774436090225564, + "grad_norm": 0.08420135080814362, + "learning_rate": 0.0002585, + "loss": 0.031, + "step": 5170 + }, + { + "epoch": 7.7894736842105265, + "grad_norm": 0.05334215238690376, + "learning_rate": 0.000259, + "loss": 0.0321, + "step": 5180 + }, + { + "epoch": 7.804511278195489, + "grad_norm": 0.08204364031553268, + "learning_rate": 0.0002595, + "loss": 0.0374, + "step": 5190 + }, + { + "epoch": 7.819548872180452, + "grad_norm": 0.0516546405851841, + "learning_rate": 0.00026000000000000003, + "loss": 0.0295, + "step": 5200 + }, + { + "epoch": 7.834586466165414, + "grad_norm": 0.1068059578537941, + "learning_rate": 0.0002605, + "loss": 0.0319, + "step": 5210 + }, + { + "epoch": 7.849624060150376, + "grad_norm": 0.07823677361011505, + "learning_rate": 0.000261, + "loss": 0.0319, + "step": 5220 + }, + { + "epoch": 7.864661654135339, + "grad_norm": 0.06938464194536209, + "learning_rate": 0.0002615, + "loss": 0.0337, + "step": 5230 + }, + { + "epoch": 7.879699248120301, + "grad_norm": 0.04994886368513107, + "learning_rate": 0.000262, + "loss": 0.0337, + "step": 5240 + }, + { + "epoch": 7.894736842105263, + "grad_norm": 0.06543327867984772, + "learning_rate": 0.00026250000000000004, + "loss": 0.0374, + "step": 5250 + }, + { + "epoch": 7.909774436090226, + "grad_norm": 0.0770590528845787, + "learning_rate": 0.000263, + "loss": 0.0303, + "step": 5260 + }, + { + "epoch": 7.924812030075188, + "grad_norm": 0.067111074924469, + "learning_rate": 0.0002635, + "loss": 0.0358, + "step": 5270 + }, + { + "epoch": 7.93984962406015, + "grad_norm": 0.06502697616815567, + "learning_rate": 0.000264, + "loss": 0.0308, + "step": 5280 + }, + { + "epoch": 7.954887218045113, + "grad_norm": 0.05524824187159538, + "learning_rate": 0.00026450000000000003, + "loss": 0.0297, + "step": 5290 + }, + { + "epoch": 7.969924812030075, + "grad_norm": 0.052187200635671616, + "learning_rate": 0.00026500000000000004, + "loss": 0.0312, + "step": 5300 + }, + { + "epoch": 7.984962406015038, + "grad_norm": 0.0663803368806839, + "learning_rate": 0.0002655, + "loss": 0.0309, + "step": 5310 + }, + { + "epoch": 8.0, + "grad_norm": 0.07302431762218475, + "learning_rate": 0.000266, + "loss": 0.0316, + "step": 5320 + }, + { + "epoch": 8.015037593984962, + "grad_norm": 0.07671147584915161, + "learning_rate": 0.0002665, + "loss": 0.0285, + "step": 5330 + }, + { + "epoch": 8.030075187969924, + "grad_norm": 0.049163371324539185, + "learning_rate": 0.00026700000000000004, + "loss": 0.0304, + "step": 5340 + }, + { + "epoch": 8.045112781954888, + "grad_norm": 0.06664051115512848, + "learning_rate": 0.0002675, + "loss": 0.0303, + "step": 5350 + }, + { + "epoch": 8.06015037593985, + "grad_norm": 0.06028413027524948, + "learning_rate": 0.000268, + "loss": 0.0288, + "step": 5360 + }, + { + "epoch": 8.075187969924812, + "grad_norm": 0.056082822382450104, + "learning_rate": 0.0002685, + "loss": 0.0287, + "step": 5370 + }, + { + "epoch": 8.090225563909774, + "grad_norm": 0.0727391242980957, + "learning_rate": 0.00026900000000000003, + "loss": 0.0314, + "step": 5380 + }, + { + "epoch": 8.105263157894736, + "grad_norm": 0.08558163791894913, + "learning_rate": 0.00026950000000000005, + "loss": 0.0275, + "step": 5390 + }, + { + "epoch": 8.1203007518797, + "grad_norm": 0.07853320240974426, + "learning_rate": 0.00027, + "loss": 0.0286, + "step": 5400 + }, + { + "epoch": 8.135338345864662, + "grad_norm": 0.06521563231945038, + "learning_rate": 0.0002705, + "loss": 0.0313, + "step": 5410 + }, + { + "epoch": 8.150375939849624, + "grad_norm": 0.053905729204416275, + "learning_rate": 0.00027100000000000003, + "loss": 0.0282, + "step": 5420 + }, + { + "epoch": 8.165413533834586, + "grad_norm": 0.06751992553472519, + "learning_rate": 0.00027150000000000004, + "loss": 0.03, + "step": 5430 + }, + { + "epoch": 8.180451127819548, + "grad_norm": 0.0752953290939331, + "learning_rate": 0.00027200000000000005, + "loss": 0.0319, + "step": 5440 + }, + { + "epoch": 8.19548872180451, + "grad_norm": 0.08498464524745941, + "learning_rate": 0.0002725, + "loss": 0.0276, + "step": 5450 + }, + { + "epoch": 8.210526315789474, + "grad_norm": 0.07531293481588364, + "learning_rate": 0.000273, + "loss": 0.0317, + "step": 5460 + }, + { + "epoch": 8.225563909774436, + "grad_norm": 0.04610944911837578, + "learning_rate": 0.00027350000000000003, + "loss": 0.0304, + "step": 5470 + }, + { + "epoch": 8.240601503759398, + "grad_norm": 0.06423097848892212, + "learning_rate": 0.00027400000000000005, + "loss": 0.0293, + "step": 5480 + }, + { + "epoch": 8.25563909774436, + "grad_norm": 0.05764598026871681, + "learning_rate": 0.0002745, + "loss": 0.0305, + "step": 5490 + }, + { + "epoch": 8.270676691729323, + "grad_norm": 0.0868816003203392, + "learning_rate": 0.000275, + "loss": 0.0297, + "step": 5500 + }, + { + "epoch": 8.285714285714286, + "grad_norm": 0.06227708235383034, + "learning_rate": 0.00027550000000000003, + "loss": 0.0285, + "step": 5510 + }, + { + "epoch": 8.300751879699249, + "grad_norm": 0.06763945519924164, + "learning_rate": 0.00027600000000000004, + "loss": 0.0316, + "step": 5520 + }, + { + "epoch": 8.31578947368421, + "grad_norm": 0.05977148562669754, + "learning_rate": 0.00027650000000000005, + "loss": 0.0251, + "step": 5530 + }, + { + "epoch": 8.330827067669173, + "grad_norm": 0.05702297389507294, + "learning_rate": 0.000277, + "loss": 0.0279, + "step": 5540 + }, + { + "epoch": 8.345864661654135, + "grad_norm": 0.07489630579948425, + "learning_rate": 0.0002775, + "loss": 0.0261, + "step": 5550 + }, + { + "epoch": 8.360902255639097, + "grad_norm": 0.07320330291986465, + "learning_rate": 0.00027800000000000004, + "loss": 0.0316, + "step": 5560 + }, + { + "epoch": 8.37593984962406, + "grad_norm": 0.08490370959043503, + "learning_rate": 0.00027850000000000005, + "loss": 0.0278, + "step": 5570 + }, + { + "epoch": 8.390977443609023, + "grad_norm": 0.12464156001806259, + "learning_rate": 0.000279, + "loss": 0.0287, + "step": 5580 + }, + { + "epoch": 8.406015037593985, + "grad_norm": 0.0525430366396904, + "learning_rate": 0.0002795, + "loss": 0.028, + "step": 5590 + }, + { + "epoch": 8.421052631578947, + "grad_norm": 0.06532420963048935, + "learning_rate": 0.00028000000000000003, + "loss": 0.0305, + "step": 5600 + }, + { + "epoch": 8.436090225563909, + "grad_norm": 0.05026834085583687, + "learning_rate": 0.00028050000000000004, + "loss": 0.0256, + "step": 5610 + }, + { + "epoch": 8.451127819548873, + "grad_norm": 0.05492403358221054, + "learning_rate": 0.00028100000000000005, + "loss": 0.0323, + "step": 5620 + }, + { + "epoch": 8.466165413533835, + "grad_norm": 0.07541689276695251, + "learning_rate": 0.00028149999999999996, + "loss": 0.0281, + "step": 5630 + }, + { + "epoch": 8.481203007518797, + "grad_norm": 0.05638613924384117, + "learning_rate": 0.00028199999999999997, + "loss": 0.0297, + "step": 5640 + }, + { + "epoch": 8.496240601503759, + "grad_norm": 0.05785512179136276, + "learning_rate": 0.0002825, + "loss": 0.0305, + "step": 5650 + }, + { + "epoch": 8.511278195488721, + "grad_norm": 0.06490733474493027, + "learning_rate": 0.000283, + "loss": 0.0313, + "step": 5660 + }, + { + "epoch": 8.526315789473685, + "grad_norm": 0.08833730965852737, + "learning_rate": 0.0002835, + "loss": 0.0266, + "step": 5670 + }, + { + "epoch": 8.541353383458647, + "grad_norm": 0.058735135942697525, + "learning_rate": 0.00028399999999999996, + "loss": 0.0276, + "step": 5680 + }, + { + "epoch": 8.556390977443609, + "grad_norm": 0.06440144777297974, + "learning_rate": 0.0002845, + "loss": 0.0287, + "step": 5690 + }, + { + "epoch": 8.571428571428571, + "grad_norm": 0.05779734626412392, + "learning_rate": 0.000285, + "loss": 0.0277, + "step": 5700 + }, + { + "epoch": 8.586466165413533, + "grad_norm": 0.05781801789999008, + "learning_rate": 0.0002855, + "loss": 0.0291, + "step": 5710 + }, + { + "epoch": 8.601503759398497, + "grad_norm": 0.060084182769060135, + "learning_rate": 0.00028599999999999996, + "loss": 0.0305, + "step": 5720 + }, + { + "epoch": 8.61654135338346, + "grad_norm": 0.06874674558639526, + "learning_rate": 0.00028649999999999997, + "loss": 0.0286, + "step": 5730 + }, + { + "epoch": 8.631578947368421, + "grad_norm": 0.06905695796012878, + "learning_rate": 0.000287, + "loss": 0.0285, + "step": 5740 + }, + { + "epoch": 8.646616541353383, + "grad_norm": 0.07849803566932678, + "learning_rate": 0.0002875, + "loss": 0.0289, + "step": 5750 + }, + { + "epoch": 8.661654135338345, + "grad_norm": 0.05800304561853409, + "learning_rate": 0.000288, + "loss": 0.0284, + "step": 5760 + }, + { + "epoch": 8.676691729323307, + "grad_norm": 0.09472599625587463, + "learning_rate": 0.00028849999999999997, + "loss": 0.0286, + "step": 5770 + }, + { + "epoch": 8.691729323308271, + "grad_norm": 0.07318698614835739, + "learning_rate": 0.000289, + "loss": 0.0296, + "step": 5780 + }, + { + "epoch": 8.706766917293233, + "grad_norm": 0.1000836119055748, + "learning_rate": 0.0002895, + "loss": 0.0314, + "step": 5790 + }, + { + "epoch": 8.721804511278195, + "grad_norm": 0.055665381252765656, + "learning_rate": 0.00029, + "loss": 0.0262, + "step": 5800 + }, + { + "epoch": 8.736842105263158, + "grad_norm": 0.08088107407093048, + "learning_rate": 0.00029049999999999996, + "loss": 0.0289, + "step": 5810 + }, + { + "epoch": 8.75187969924812, + "grad_norm": 0.06867188960313797, + "learning_rate": 0.00029099999999999997, + "loss": 0.0277, + "step": 5820 + }, + { + "epoch": 8.766917293233083, + "grad_norm": 0.06158306449651718, + "learning_rate": 0.0002915, + "loss": 0.0266, + "step": 5830 + }, + { + "epoch": 8.781954887218046, + "grad_norm": 0.053368985652923584, + "learning_rate": 0.000292, + "loss": 0.028, + "step": 5840 + }, + { + "epoch": 8.796992481203008, + "grad_norm": 0.05575254559516907, + "learning_rate": 0.0002925, + "loss": 0.0311, + "step": 5850 + }, + { + "epoch": 8.81203007518797, + "grad_norm": 0.06802136451005936, + "learning_rate": 0.00029299999999999997, + "loss": 0.028, + "step": 5860 + }, + { + "epoch": 8.827067669172932, + "grad_norm": 0.06745045632123947, + "learning_rate": 0.0002935, + "loss": 0.0264, + "step": 5870 + }, + { + "epoch": 8.842105263157894, + "grad_norm": 0.1076737642288208, + "learning_rate": 0.000294, + "loss": 0.0287, + "step": 5880 + }, + { + "epoch": 8.857142857142858, + "grad_norm": 0.08280255645513535, + "learning_rate": 0.0002945, + "loss": 0.0301, + "step": 5890 + }, + { + "epoch": 8.87218045112782, + "grad_norm": 0.08360929042100906, + "learning_rate": 0.000295, + "loss": 0.0306, + "step": 5900 + }, + { + "epoch": 8.887218045112782, + "grad_norm": 0.04834266006946564, + "learning_rate": 0.00029549999999999997, + "loss": 0.0295, + "step": 5910 + }, + { + "epoch": 8.902255639097744, + "grad_norm": 0.05208633095026016, + "learning_rate": 0.000296, + "loss": 0.0273, + "step": 5920 + }, + { + "epoch": 8.917293233082706, + "grad_norm": 0.07123664766550064, + "learning_rate": 0.0002965, + "loss": 0.0272, + "step": 5930 + }, + { + "epoch": 8.93233082706767, + "grad_norm": 0.12283273041248322, + "learning_rate": 0.000297, + "loss": 0.0266, + "step": 5940 + }, + { + "epoch": 8.947368421052632, + "grad_norm": 0.056478142738342285, + "learning_rate": 0.00029749999999999997, + "loss": 0.0303, + "step": 5950 + }, + { + "epoch": 8.962406015037594, + "grad_norm": 0.05896707996726036, + "learning_rate": 0.000298, + "loss": 0.0286, + "step": 5960 + }, + { + "epoch": 8.977443609022556, + "grad_norm": 0.07083549350500107, + "learning_rate": 0.0002985, + "loss": 0.027, + "step": 5970 + }, + { + "epoch": 8.992481203007518, + "grad_norm": 0.04566838964819908, + "learning_rate": 0.000299, + "loss": 0.0272, + "step": 5980 + }, + { + "epoch": 9.007518796992482, + "grad_norm": 0.07623133808374405, + "learning_rate": 0.0002995, + "loss": 0.0263, + "step": 5990 + }, + { + "epoch": 9.022556390977444, + "grad_norm": 0.05661000683903694, + "learning_rate": 0.0003, + "loss": 0.0256, + "step": 6000 + }, + { + "epoch": 9.022556390977444, + "eval_cer": 0.02089932661352295, + "eval_loss": 0.049986857920885086, + "eval_runtime": 162.1843, + "eval_samples_per_second": 98.82, + "eval_steps_per_second": 0.777, + "eval_wer": 0.06893570636546449, + "step": 6000 + }, + { + "epoch": 9.037593984962406, + "grad_norm": 0.0646754801273346, + "learning_rate": 0.0003005, + "loss": 0.0248, + "step": 6010 + }, + { + "epoch": 9.052631578947368, + "grad_norm": 0.05372212082147598, + "learning_rate": 0.000301, + "loss": 0.0268, + "step": 6020 + }, + { + "epoch": 9.06766917293233, + "grad_norm": 0.06253086775541306, + "learning_rate": 0.0003015, + "loss": 0.0259, + "step": 6030 + }, + { + "epoch": 9.082706766917294, + "grad_norm": 0.06135258078575134, + "learning_rate": 0.000302, + "loss": 0.024, + "step": 6040 + }, + { + "epoch": 9.097744360902256, + "grad_norm": 0.06333441287279129, + "learning_rate": 0.0003025, + "loss": 0.025, + "step": 6050 + }, + { + "epoch": 9.112781954887218, + "grad_norm": 0.06290154159069061, + "learning_rate": 0.000303, + "loss": 0.0222, + "step": 6060 + }, + { + "epoch": 9.12781954887218, + "grad_norm": 0.0728684589266777, + "learning_rate": 0.0003035, + "loss": 0.0244, + "step": 6070 + }, + { + "epoch": 9.142857142857142, + "grad_norm": 0.058562543243169785, + "learning_rate": 0.000304, + "loss": 0.0252, + "step": 6080 + }, + { + "epoch": 9.157894736842104, + "grad_norm": 0.0583585686981678, + "learning_rate": 0.0003045, + "loss": 0.028, + "step": 6090 + }, + { + "epoch": 9.172932330827068, + "grad_norm": 0.069390207529068, + "learning_rate": 0.000305, + "loss": 0.0219, + "step": 6100 + }, + { + "epoch": 9.18796992481203, + "grad_norm": 0.10119915008544922, + "learning_rate": 0.0003055, + "loss": 0.0242, + "step": 6110 + }, + { + "epoch": 9.203007518796992, + "grad_norm": 0.052722346037626266, + "learning_rate": 0.000306, + "loss": 0.0241, + "step": 6120 + }, + { + "epoch": 9.218045112781954, + "grad_norm": 0.056572381407022476, + "learning_rate": 0.0003065, + "loss": 0.0228, + "step": 6130 + }, + { + "epoch": 9.233082706766917, + "grad_norm": 0.062166132032871246, + "learning_rate": 0.000307, + "loss": 0.0237, + "step": 6140 + }, + { + "epoch": 9.24812030075188, + "grad_norm": 0.10071983188390732, + "learning_rate": 0.0003075, + "loss": 0.0275, + "step": 6150 + }, + { + "epoch": 9.263157894736842, + "grad_norm": 0.0663413554430008, + "learning_rate": 0.000308, + "loss": 0.0245, + "step": 6160 + }, + { + "epoch": 9.278195488721805, + "grad_norm": 0.05554046481847763, + "learning_rate": 0.0003085, + "loss": 0.0236, + "step": 6170 + }, + { + "epoch": 9.293233082706767, + "grad_norm": 0.054320912808179855, + "learning_rate": 0.00030900000000000003, + "loss": 0.0267, + "step": 6180 + }, + { + "epoch": 9.308270676691729, + "grad_norm": 0.06805162131786346, + "learning_rate": 0.0003095, + "loss": 0.0258, + "step": 6190 + }, + { + "epoch": 9.323308270676693, + "grad_norm": 0.060956936329603195, + "learning_rate": 0.00031, + "loss": 0.0254, + "step": 6200 + }, + { + "epoch": 9.338345864661655, + "grad_norm": 0.06719738990068436, + "learning_rate": 0.0003105, + "loss": 0.0243, + "step": 6210 + }, + { + "epoch": 9.353383458646617, + "grad_norm": 0.09574072808027267, + "learning_rate": 0.000311, + "loss": 0.0265, + "step": 6220 + }, + { + "epoch": 9.368421052631579, + "grad_norm": 0.06389805674552917, + "learning_rate": 0.0003115, + "loss": 0.0259, + "step": 6230 + }, + { + "epoch": 9.38345864661654, + "grad_norm": 0.08743830025196075, + "learning_rate": 0.000312, + "loss": 0.0248, + "step": 6240 + }, + { + "epoch": 9.398496240601503, + "grad_norm": 0.06206941977143288, + "learning_rate": 0.0003125, + "loss": 0.0275, + "step": 6250 + }, + { + "epoch": 9.413533834586467, + "grad_norm": 0.07709711790084839, + "learning_rate": 0.000313, + "loss": 0.0281, + "step": 6260 + }, + { + "epoch": 9.428571428571429, + "grad_norm": 0.06280162930488586, + "learning_rate": 0.00031350000000000003, + "loss": 0.0261, + "step": 6270 + }, + { + "epoch": 9.443609022556391, + "grad_norm": 0.064295694231987, + "learning_rate": 0.000314, + "loss": 0.025, + "step": 6280 + }, + { + "epoch": 9.458646616541353, + "grad_norm": 0.05373825505375862, + "learning_rate": 0.0003145, + "loss": 0.0223, + "step": 6290 + }, + { + "epoch": 9.473684210526315, + "grad_norm": 0.06735815107822418, + "learning_rate": 0.000315, + "loss": 0.0236, + "step": 6300 + }, + { + "epoch": 9.488721804511279, + "grad_norm": 0.06319549679756165, + "learning_rate": 0.0003155, + "loss": 0.0233, + "step": 6310 + }, + { + "epoch": 9.503759398496241, + "grad_norm": 0.05680960416793823, + "learning_rate": 0.000316, + "loss": 0.0251, + "step": 6320 + }, + { + "epoch": 9.518796992481203, + "grad_norm": 0.06428761035203934, + "learning_rate": 0.0003165, + "loss": 0.0236, + "step": 6330 + }, + { + "epoch": 9.533834586466165, + "grad_norm": 0.08156956732273102, + "learning_rate": 0.000317, + "loss": 0.0253, + "step": 6340 + }, + { + "epoch": 9.548872180451127, + "grad_norm": 0.08069149404764175, + "learning_rate": 0.0003175, + "loss": 0.0239, + "step": 6350 + }, + { + "epoch": 9.563909774436091, + "grad_norm": 0.10007152706384659, + "learning_rate": 0.00031800000000000003, + "loss": 0.0251, + "step": 6360 + }, + { + "epoch": 9.578947368421053, + "grad_norm": 0.07681086659431458, + "learning_rate": 0.0003185, + "loss": 0.0284, + "step": 6370 + }, + { + "epoch": 9.593984962406015, + "grad_norm": 0.09057555347681046, + "learning_rate": 0.000319, + "loss": 0.0255, + "step": 6380 + }, + { + "epoch": 9.609022556390977, + "grad_norm": 0.05278893932700157, + "learning_rate": 0.0003195, + "loss": 0.0271, + "step": 6390 + }, + { + "epoch": 9.62406015037594, + "grad_norm": 0.0719471201300621, + "learning_rate": 0.00032, + "loss": 0.0224, + "step": 6400 + }, + { + "epoch": 9.639097744360903, + "grad_norm": 0.06820385903120041, + "learning_rate": 0.00032050000000000004, + "loss": 0.0268, + "step": 6410 + }, + { + "epoch": 9.654135338345865, + "grad_norm": 0.06903686374425888, + "learning_rate": 0.000321, + "loss": 0.0271, + "step": 6420 + }, + { + "epoch": 9.669172932330827, + "grad_norm": 0.05220801755785942, + "learning_rate": 0.0003215, + "loss": 0.0242, + "step": 6430 + }, + { + "epoch": 9.68421052631579, + "grad_norm": 0.057818423956632614, + "learning_rate": 0.000322, + "loss": 0.0275, + "step": 6440 + }, + { + "epoch": 9.699248120300751, + "grad_norm": 0.05058760941028595, + "learning_rate": 0.00032250000000000003, + "loss": 0.0233, + "step": 6450 + }, + { + "epoch": 9.714285714285714, + "grad_norm": 0.07405845075845718, + "learning_rate": 0.000323, + "loss": 0.0239, + "step": 6460 + }, + { + "epoch": 9.729323308270677, + "grad_norm": 0.0832362100481987, + "learning_rate": 0.0003235, + "loss": 0.0246, + "step": 6470 + }, + { + "epoch": 9.74436090225564, + "grad_norm": 0.08253604918718338, + "learning_rate": 0.000324, + "loss": 0.0226, + "step": 6480 + }, + { + "epoch": 9.759398496240602, + "grad_norm": 0.0561661422252655, + "learning_rate": 0.00032450000000000003, + "loss": 0.024, + "step": 6490 + }, + { + "epoch": 9.774436090225564, + "grad_norm": 0.05794068053364754, + "learning_rate": 0.00032500000000000004, + "loss": 0.0228, + "step": 6500 + }, + { + "epoch": 9.789473684210526, + "grad_norm": 0.04981200397014618, + "learning_rate": 0.0003255, + "loss": 0.0259, + "step": 6510 + }, + { + "epoch": 9.80451127819549, + "grad_norm": 0.05791035667061806, + "learning_rate": 0.000326, + "loss": 0.021, + "step": 6520 + }, + { + "epoch": 9.819548872180452, + "grad_norm": 0.07462535798549652, + "learning_rate": 0.0003265, + "loss": 0.0213, + "step": 6530 + }, + { + "epoch": 9.834586466165414, + "grad_norm": 0.0686153918504715, + "learning_rate": 0.00032700000000000003, + "loss": 0.0241, + "step": 6540 + }, + { + "epoch": 9.849624060150376, + "grad_norm": 0.056861359626054764, + "learning_rate": 0.00032750000000000005, + "loss": 0.0242, + "step": 6550 + }, + { + "epoch": 9.864661654135338, + "grad_norm": 0.07829906046390533, + "learning_rate": 0.000328, + "loss": 0.0226, + "step": 6560 + }, + { + "epoch": 9.8796992481203, + "grad_norm": 0.08958689868450165, + "learning_rate": 0.0003285, + "loss": 0.0254, + "step": 6570 + }, + { + "epoch": 9.894736842105264, + "grad_norm": 0.08215101063251495, + "learning_rate": 0.00032900000000000003, + "loss": 0.0269, + "step": 6580 + }, + { + "epoch": 9.909774436090226, + "grad_norm": 0.22095702588558197, + "learning_rate": 0.00032950000000000004, + "loss": 0.0249, + "step": 6590 + }, + { + "epoch": 9.924812030075188, + "grad_norm": 0.07950076460838318, + "learning_rate": 0.00033, + "loss": 0.023, + "step": 6600 + }, + { + "epoch": 9.93984962406015, + "grad_norm": 0.065628282725811, + "learning_rate": 0.0003305, + "loss": 0.0252, + "step": 6610 + }, + { + "epoch": 9.954887218045112, + "grad_norm": 0.059811074286699295, + "learning_rate": 0.000331, + "loss": 0.0236, + "step": 6620 + }, + { + "epoch": 9.969924812030076, + "grad_norm": 0.07975172996520996, + "learning_rate": 0.00033150000000000003, + "loss": 0.0254, + "step": 6630 + }, + { + "epoch": 9.984962406015038, + "grad_norm": 0.05912427976727486, + "learning_rate": 0.00033200000000000005, + "loss": 0.027, + "step": 6640 + }, + { + "epoch": 10.0, + "grad_norm": 0.0513499453663826, + "learning_rate": 0.0003325, + "loss": 0.025, + "step": 6650 + }, + { + "epoch": 10.015037593984962, + "grad_norm": 0.06787694990634918, + "learning_rate": 0.000333, + "loss": 0.0215, + "step": 6660 + }, + { + "epoch": 10.030075187969924, + "grad_norm": 0.04847653955221176, + "learning_rate": 0.00033350000000000003, + "loss": 0.0188, + "step": 6670 + }, + { + "epoch": 10.045112781954888, + "grad_norm": 0.07146945595741272, + "learning_rate": 0.00033400000000000004, + "loss": 0.0207, + "step": 6680 + }, + { + "epoch": 10.06015037593985, + "grad_norm": 0.06184151768684387, + "learning_rate": 0.00033450000000000005, + "loss": 0.0204, + "step": 6690 + }, + { + "epoch": 10.075187969924812, + "grad_norm": 0.05831315368413925, + "learning_rate": 0.000335, + "loss": 0.0222, + "step": 6700 + }, + { + "epoch": 10.090225563909774, + "grad_norm": 0.08657033741474152, + "learning_rate": 0.0003355, + "loss": 0.024, + "step": 6710 + }, + { + "epoch": 10.105263157894736, + "grad_norm": 0.05963308364152908, + "learning_rate": 0.00033600000000000004, + "loss": 0.0207, + "step": 6720 + }, + { + "epoch": 10.1203007518797, + "grad_norm": 0.0681181475520134, + "learning_rate": 0.00033650000000000005, + "loss": 0.0218, + "step": 6730 + }, + { + "epoch": 10.135338345864662, + "grad_norm": 0.06210342422127724, + "learning_rate": 0.000337, + "loss": 0.0203, + "step": 6740 + }, + { + "epoch": 10.150375939849624, + "grad_norm": 0.06268936395645142, + "learning_rate": 0.0003375, + "loss": 0.0211, + "step": 6750 + }, + { + "epoch": 10.165413533834586, + "grad_norm": 0.05569452419877052, + "learning_rate": 0.00033800000000000003, + "loss": 0.0223, + "step": 6760 + }, + { + "epoch": 10.180451127819548, + "grad_norm": 0.08302775770425797, + "learning_rate": 0.00033850000000000004, + "loss": 0.0208, + "step": 6770 + }, + { + "epoch": 10.19548872180451, + "grad_norm": 0.054263804107904434, + "learning_rate": 0.00033900000000000005, + "loss": 0.0198, + "step": 6780 + }, + { + "epoch": 10.210526315789474, + "grad_norm": 0.0629635825753212, + "learning_rate": 0.0003395, + "loss": 0.0217, + "step": 6790 + }, + { + "epoch": 10.225563909774436, + "grad_norm": 0.0553508996963501, + "learning_rate": 0.00034, + "loss": 0.0218, + "step": 6800 + }, + { + "epoch": 10.240601503759398, + "grad_norm": 0.04902573302388191, + "learning_rate": 0.00034050000000000004, + "loss": 0.022, + "step": 6810 + }, + { + "epoch": 10.25563909774436, + "grad_norm": 0.06371958553791046, + "learning_rate": 0.00034100000000000005, + "loss": 0.02, + "step": 6820 + }, + { + "epoch": 10.270676691729323, + "grad_norm": 0.0682794451713562, + "learning_rate": 0.0003415, + "loss": 0.0195, + "step": 6830 + }, + { + "epoch": 10.285714285714286, + "grad_norm": 0.07929519563913345, + "learning_rate": 0.000342, + "loss": 0.0212, + "step": 6840 + }, + { + "epoch": 10.300751879699249, + "grad_norm": 0.04800889268517494, + "learning_rate": 0.00034250000000000003, + "loss": 0.0189, + "step": 6850 + }, + { + "epoch": 10.31578947368421, + "grad_norm": 0.0675400048494339, + "learning_rate": 0.00034300000000000004, + "loss": 0.0223, + "step": 6860 + }, + { + "epoch": 10.330827067669173, + "grad_norm": 0.11442269384860992, + "learning_rate": 0.00034350000000000006, + "loss": 0.0236, + "step": 6870 + }, + { + "epoch": 10.345864661654135, + "grad_norm": 0.05798109620809555, + "learning_rate": 0.00034399999999999996, + "loss": 0.021, + "step": 6880 + }, + { + "epoch": 10.360902255639097, + "grad_norm": 0.06602201610803604, + "learning_rate": 0.00034449999999999997, + "loss": 0.0257, + "step": 6890 + }, + { + "epoch": 10.37593984962406, + "grad_norm": 0.06187829747796059, + "learning_rate": 0.000345, + "loss": 0.0199, + "step": 6900 + }, + { + "epoch": 10.390977443609023, + "grad_norm": 0.055874817073345184, + "learning_rate": 0.0003455, + "loss": 0.0226, + "step": 6910 + }, + { + "epoch": 10.406015037593985, + "grad_norm": 0.04841848462820053, + "learning_rate": 0.000346, + "loss": 0.0228, + "step": 6920 + }, + { + "epoch": 10.421052631578947, + "grad_norm": 0.07154963165521622, + "learning_rate": 0.00034649999999999997, + "loss": 0.0216, + "step": 6930 + }, + { + "epoch": 10.436090225563909, + "grad_norm": 0.07345172762870789, + "learning_rate": 0.000347, + "loss": 0.0207, + "step": 6940 + }, + { + "epoch": 10.451127819548873, + "grad_norm": 0.07390587031841278, + "learning_rate": 0.0003475, + "loss": 0.0237, + "step": 6950 + }, + { + "epoch": 10.466165413533835, + "grad_norm": 0.08680398762226105, + "learning_rate": 0.000348, + "loss": 0.024, + "step": 6960 + }, + { + "epoch": 10.481203007518797, + "grad_norm": 0.04506571590900421, + "learning_rate": 0.00034849999999999996, + "loss": 0.0218, + "step": 6970 + }, + { + "epoch": 10.496240601503759, + "grad_norm": 0.06755760312080383, + "learning_rate": 0.00034899999999999997, + "loss": 0.0217, + "step": 6980 + }, + { + "epoch": 10.511278195488721, + "grad_norm": 0.0867224857211113, + "learning_rate": 0.0003495, + "loss": 0.0204, + "step": 6990 + }, + { + "epoch": 10.526315789473685, + "grad_norm": 0.12629158794879913, + "learning_rate": 0.00035, + "loss": 0.0209, + "step": 7000 + }, + { + "epoch": 10.526315789473685, + "eval_cer": 0.020870764472504858, + "eval_loss": 0.053246669471263885, + "eval_runtime": 160.7183, + "eval_samples_per_second": 99.721, + "eval_steps_per_second": 0.784, + "eval_wer": 0.07069292788813923, + "step": 7000 + }, + { + "epoch": 10.541353383458647, + "grad_norm": 0.07012065500020981, + "learning_rate": 0.0003505, + "loss": 0.0213, + "step": 7010 + }, + { + "epoch": 10.556390977443609, + "grad_norm": 0.06177555397152901, + "learning_rate": 0.00035099999999999997, + "loss": 0.0223, + "step": 7020 + }, + { + "epoch": 10.571428571428571, + "grad_norm": 0.039833322167396545, + "learning_rate": 0.0003515, + "loss": 0.0193, + "step": 7030 + }, + { + "epoch": 10.586466165413533, + "grad_norm": 0.04750439524650574, + "learning_rate": 0.000352, + "loss": 0.0188, + "step": 7040 + }, + { + "epoch": 10.601503759398497, + "grad_norm": 0.05063023790717125, + "learning_rate": 0.0003525, + "loss": 0.021, + "step": 7050 + }, + { + "epoch": 10.61654135338346, + "grad_norm": 0.06804497539997101, + "learning_rate": 0.00035299999999999996, + "loss": 0.0225, + "step": 7060 + }, + { + "epoch": 10.631578947368421, + "grad_norm": 0.057973917573690414, + "learning_rate": 0.0003535, + "loss": 0.022, + "step": 7070 + }, + { + "epoch": 10.646616541353383, + "grad_norm": 0.05432054027915001, + "learning_rate": 0.000354, + "loss": 0.0232, + "step": 7080 + }, + { + "epoch": 10.661654135338345, + "grad_norm": 0.0589754693210125, + "learning_rate": 0.0003545, + "loss": 0.0203, + "step": 7090 + }, + { + "epoch": 10.676691729323307, + "grad_norm": 0.05649235472083092, + "learning_rate": 0.000355, + "loss": 0.024, + "step": 7100 + }, + { + "epoch": 10.691729323308271, + "grad_norm": 0.07018685340881348, + "learning_rate": 0.00035549999999999997, + "loss": 0.0208, + "step": 7110 + }, + { + "epoch": 10.706766917293233, + "grad_norm": 0.06444822251796722, + "learning_rate": 0.000356, + "loss": 0.0218, + "step": 7120 + }, + { + "epoch": 10.721804511278195, + "grad_norm": 0.0869486927986145, + "learning_rate": 0.0003565, + "loss": 0.0236, + "step": 7130 + }, + { + "epoch": 10.736842105263158, + "grad_norm": 0.047450754791498184, + "learning_rate": 0.000357, + "loss": 0.0214, + "step": 7140 + }, + { + "epoch": 10.75187969924812, + "grad_norm": 0.10121061652898788, + "learning_rate": 0.0003575, + "loss": 0.0189, + "step": 7150 + }, + { + "epoch": 10.766917293233083, + "grad_norm": 0.045771099627017975, + "learning_rate": 0.000358, + "loss": 0.019, + "step": 7160 + }, + { + "epoch": 10.781954887218046, + "grad_norm": 0.07388034462928772, + "learning_rate": 0.0003585, + "loss": 0.0213, + "step": 7170 + }, + { + "epoch": 10.796992481203008, + "grad_norm": 0.053388796746730804, + "learning_rate": 0.000359, + "loss": 0.0191, + "step": 7180 + }, + { + "epoch": 10.81203007518797, + "grad_norm": 0.056514590978622437, + "learning_rate": 0.0003595, + "loss": 0.0197, + "step": 7190 + }, + { + "epoch": 10.827067669172932, + "grad_norm": 0.05402562394738197, + "learning_rate": 0.00035999999999999997, + "loss": 0.0208, + "step": 7200 + }, + { + "epoch": 10.842105263157894, + "grad_norm": 0.06072898209095001, + "learning_rate": 0.0003605, + "loss": 0.0228, + "step": 7210 + }, + { + "epoch": 10.857142857142858, + "grad_norm": 0.06615385413169861, + "learning_rate": 0.000361, + "loss": 0.0211, + "step": 7220 + }, + { + "epoch": 10.87218045112782, + "grad_norm": 0.052466265857219696, + "learning_rate": 0.0003615, + "loss": 0.0222, + "step": 7230 + }, + { + "epoch": 10.887218045112782, + "grad_norm": 0.06466560810804367, + "learning_rate": 0.000362, + "loss": 0.0206, + "step": 7240 + }, + { + "epoch": 10.902255639097744, + "grad_norm": 0.08312542736530304, + "learning_rate": 0.0003625, + "loss": 0.0188, + "step": 7250 + }, + { + "epoch": 10.917293233082706, + "grad_norm": 0.05996193364262581, + "learning_rate": 0.000363, + "loss": 0.0204, + "step": 7260 + }, + { + "epoch": 10.93233082706767, + "grad_norm": 0.13046693801879883, + "learning_rate": 0.0003635, + "loss": 0.0219, + "step": 7270 + }, + { + "epoch": 10.947368421052632, + "grad_norm": 0.06853601336479187, + "learning_rate": 0.000364, + "loss": 0.0195, + "step": 7280 + }, + { + "epoch": 10.962406015037594, + "grad_norm": 0.11525923758745193, + "learning_rate": 0.0003645, + "loss": 0.0218, + "step": 7290 + }, + { + "epoch": 10.977443609022556, + "grad_norm": 0.05028758570551872, + "learning_rate": 0.000365, + "loss": 0.0175, + "step": 7300 + }, + { + "epoch": 10.992481203007518, + "grad_norm": 0.04650198295712471, + "learning_rate": 0.0003655, + "loss": 0.0206, + "step": 7310 + }, + { + "epoch": 11.007518796992482, + "grad_norm": 0.04972374066710472, + "learning_rate": 0.000366, + "loss": 0.0177, + "step": 7320 + }, + { + "epoch": 11.022556390977444, + "grad_norm": 0.05673589929938316, + "learning_rate": 0.0003665, + "loss": 0.0177, + "step": 7330 + }, + { + "epoch": 11.037593984962406, + "grad_norm": 0.07219090312719345, + "learning_rate": 0.000367, + "loss": 0.0167, + "step": 7340 + }, + { + "epoch": 11.052631578947368, + "grad_norm": 0.10747040808200836, + "learning_rate": 0.0003675, + "loss": 0.0173, + "step": 7350 + }, + { + "epoch": 11.06766917293233, + "grad_norm": 0.059174779802560806, + "learning_rate": 0.000368, + "loss": 0.0167, + "step": 7360 + }, + { + "epoch": 11.082706766917294, + "grad_norm": 0.061220861971378326, + "learning_rate": 0.0003685, + "loss": 0.0158, + "step": 7370 + }, + { + "epoch": 11.097744360902256, + "grad_norm": 0.15846867859363556, + "learning_rate": 0.000369, + "loss": 0.0183, + "step": 7380 + }, + { + "epoch": 11.112781954887218, + "grad_norm": 0.05168516933917999, + "learning_rate": 0.0003695, + "loss": 0.017, + "step": 7390 + }, + { + "epoch": 11.12781954887218, + "grad_norm": 0.0534416064620018, + "learning_rate": 0.00037, + "loss": 0.0189, + "step": 7400 + }, + { + "epoch": 11.142857142857142, + "grad_norm": 0.06253839284181595, + "learning_rate": 0.0003705, + "loss": 0.0169, + "step": 7410 + }, + { + "epoch": 11.157894736842104, + "grad_norm": 0.06972511112689972, + "learning_rate": 0.000371, + "loss": 0.0183, + "step": 7420 + }, + { + "epoch": 11.172932330827068, + "grad_norm": 0.0551367923617363, + "learning_rate": 0.00037150000000000003, + "loss": 0.0182, + "step": 7430 + }, + { + "epoch": 11.18796992481203, + "grad_norm": 0.0604320727288723, + "learning_rate": 0.000372, + "loss": 0.0171, + "step": 7440 + }, + { + "epoch": 11.203007518796992, + "grad_norm": 0.08369924128055573, + "learning_rate": 0.0003725, + "loss": 0.0178, + "step": 7450 + }, + { + "epoch": 11.218045112781954, + "grad_norm": 0.0639827772974968, + "learning_rate": 0.000373, + "loss": 0.018, + "step": 7460 + }, + { + "epoch": 11.233082706766917, + "grad_norm": 0.0607273131608963, + "learning_rate": 0.0003735, + "loss": 0.0176, + "step": 7470 + }, + { + "epoch": 11.24812030075188, + "grad_norm": 0.08093287795782089, + "learning_rate": 0.000374, + "loss": 0.0175, + "step": 7480 + }, + { + "epoch": 11.263157894736842, + "grad_norm": 0.12471053749322891, + "learning_rate": 0.0003745, + "loss": 0.0203, + "step": 7490 + }, + { + "epoch": 11.278195488721805, + "grad_norm": 0.0670183002948761, + "learning_rate": 0.000375, + "loss": 0.0191, + "step": 7500 + }, + { + "epoch": 11.293233082706767, + "grad_norm": 0.05259780213236809, + "learning_rate": 0.0003755, + "loss": 0.0163, + "step": 7510 + }, + { + "epoch": 11.308270676691729, + "grad_norm": 0.05612128973007202, + "learning_rate": 0.00037600000000000003, + "loss": 0.018, + "step": 7520 + }, + { + "epoch": 11.323308270676693, + "grad_norm": 0.06715438514947891, + "learning_rate": 0.0003765, + "loss": 0.0181, + "step": 7530 + }, + { + "epoch": 11.338345864661655, + "grad_norm": 0.0563427098095417, + "learning_rate": 0.000377, + "loss": 0.0165, + "step": 7540 + }, + { + "epoch": 11.353383458646617, + "grad_norm": 0.06310924887657166, + "learning_rate": 0.0003775, + "loss": 0.0194, + "step": 7550 + }, + { + "epoch": 11.368421052631579, + "grad_norm": 0.06464703381061554, + "learning_rate": 0.000378, + "loss": 0.0191, + "step": 7560 + }, + { + "epoch": 11.38345864661654, + "grad_norm": 0.05543947219848633, + "learning_rate": 0.0003785, + "loss": 0.0178, + "step": 7570 + }, + { + "epoch": 11.398496240601503, + "grad_norm": 0.0875401720404625, + "learning_rate": 0.000379, + "loss": 0.019, + "step": 7580 + }, + { + "epoch": 11.413533834586467, + "grad_norm": 0.07750091701745987, + "learning_rate": 0.0003795, + "loss": 0.0184, + "step": 7590 + }, + { + "epoch": 11.428571428571429, + "grad_norm": 0.07588869333267212, + "learning_rate": 0.00038, + "loss": 0.0176, + "step": 7600 + }, + { + "epoch": 11.443609022556391, + "grad_norm": 0.06366249918937683, + "learning_rate": 0.00038050000000000003, + "loss": 0.0172, + "step": 7610 + }, + { + "epoch": 11.458646616541353, + "grad_norm": 0.06616021692752838, + "learning_rate": 0.000381, + "loss": 0.016, + "step": 7620 + }, + { + "epoch": 11.473684210526315, + "grad_norm": 0.06909584999084473, + "learning_rate": 0.0003815, + "loss": 0.0163, + "step": 7630 + }, + { + "epoch": 11.488721804511279, + "grad_norm": 0.07664395868778229, + "learning_rate": 0.000382, + "loss": 0.0188, + "step": 7640 + }, + { + "epoch": 11.503759398496241, + "grad_norm": 0.05825803801417351, + "learning_rate": 0.00038250000000000003, + "loss": 0.0185, + "step": 7650 + }, + { + "epoch": 11.518796992481203, + "grad_norm": 0.0670914426445961, + "learning_rate": 0.00038300000000000004, + "loss": 0.017, + "step": 7660 + }, + { + "epoch": 11.533834586466165, + "grad_norm": 0.058599673211574554, + "learning_rate": 0.0003835, + "loss": 0.0189, + "step": 7670 + }, + { + "epoch": 11.548872180451127, + "grad_norm": 0.057967573404312134, + "learning_rate": 0.000384, + "loss": 0.0188, + "step": 7680 + }, + { + "epoch": 11.563909774436091, + "grad_norm": 0.06156366318464279, + "learning_rate": 0.0003845, + "loss": 0.0189, + "step": 7690 + }, + { + "epoch": 11.578947368421053, + "grad_norm": 0.05368712916970253, + "learning_rate": 0.00038500000000000003, + "loss": 0.0188, + "step": 7700 + }, + { + "epoch": 11.593984962406015, + "grad_norm": 0.04518837109208107, + "learning_rate": 0.0003855, + "loss": 0.0166, + "step": 7710 + }, + { + "epoch": 11.609022556390977, + "grad_norm": 0.05398552492260933, + "learning_rate": 0.000386, + "loss": 0.0169, + "step": 7720 + }, + { + "epoch": 11.62406015037594, + "grad_norm": 0.053229644894599915, + "learning_rate": 0.0003865, + "loss": 0.0194, + "step": 7730 + }, + { + "epoch": 11.639097744360903, + "grad_norm": 0.06366915255784988, + "learning_rate": 0.00038700000000000003, + "loss": 0.0193, + "step": 7740 + }, + { + "epoch": 11.654135338345865, + "grad_norm": 0.06430919468402863, + "learning_rate": 0.00038750000000000004, + "loss": 0.0198, + "step": 7750 + }, + { + "epoch": 11.669172932330827, + "grad_norm": 0.05303851142525673, + "learning_rate": 0.000388, + "loss": 0.0182, + "step": 7760 + }, + { + "epoch": 11.68421052631579, + "grad_norm": 0.10466258972883224, + "learning_rate": 0.0003885, + "loss": 0.0175, + "step": 7770 + }, + { + "epoch": 11.699248120300751, + "grad_norm": 0.0541066899895668, + "learning_rate": 0.000389, + "loss": 0.021, + "step": 7780 + }, + { + "epoch": 11.714285714285714, + "grad_norm": 0.056870460510253906, + "learning_rate": 0.00038950000000000003, + "loss": 0.0165, + "step": 7790 + }, + { + "epoch": 11.729323308270677, + "grad_norm": 0.057444293051958084, + "learning_rate": 0.00039000000000000005, + "loss": 0.0203, + "step": 7800 + }, + { + "epoch": 11.74436090225564, + "grad_norm": 0.1251274049282074, + "learning_rate": 0.0003905, + "loss": 0.0188, + "step": 7810 + }, + { + "epoch": 11.759398496240602, + "grad_norm": 0.07483568042516708, + "learning_rate": 0.000391, + "loss": 0.0182, + "step": 7820 + }, + { + "epoch": 11.774436090225564, + "grad_norm": 0.09843656420707703, + "learning_rate": 0.00039150000000000003, + "loss": 0.0171, + "step": 7830 + }, + { + "epoch": 11.789473684210526, + "grad_norm": 0.04850930720567703, + "learning_rate": 0.00039200000000000004, + "loss": 0.0183, + "step": 7840 + }, + { + "epoch": 11.80451127819549, + "grad_norm": 0.08739500492811203, + "learning_rate": 0.0003925, + "loss": 0.0178, + "step": 7850 + }, + { + "epoch": 11.819548872180452, + "grad_norm": 0.09930869191884995, + "learning_rate": 0.000393, + "loss": 0.0192, + "step": 7860 + }, + { + "epoch": 11.834586466165414, + "grad_norm": 0.0519610270857811, + "learning_rate": 0.0003935, + "loss": 0.0171, + "step": 7870 + }, + { + "epoch": 11.849624060150376, + "grad_norm": 0.05103062465786934, + "learning_rate": 0.00039400000000000004, + "loss": 0.0193, + "step": 7880 + }, + { + "epoch": 11.864661654135338, + "grad_norm": 0.04617317020893097, + "learning_rate": 0.00039450000000000005, + "loss": 0.0159, + "step": 7890 + }, + { + "epoch": 11.8796992481203, + "grad_norm": 0.058170609176158905, + "learning_rate": 0.000395, + "loss": 0.0177, + "step": 7900 + }, + { + "epoch": 11.894736842105264, + "grad_norm": 0.053608302026987076, + "learning_rate": 0.0003955, + "loss": 0.0167, + "step": 7910 + }, + { + "epoch": 11.909774436090226, + "grad_norm": 0.1096930280327797, + "learning_rate": 0.00039600000000000003, + "loss": 0.016, + "step": 7920 + }, + { + "epoch": 11.924812030075188, + "grad_norm": 0.0591185986995697, + "learning_rate": 0.00039650000000000004, + "loss": 0.0189, + "step": 7930 + }, + { + "epoch": 11.93984962406015, + "grad_norm": 0.08197195082902908, + "learning_rate": 0.00039700000000000005, + "loss": 0.0183, + "step": 7940 + }, + { + "epoch": 11.954887218045112, + "grad_norm": 0.050988320261240005, + "learning_rate": 0.0003975, + "loss": 0.0154, + "step": 7950 + }, + { + "epoch": 11.969924812030076, + "grad_norm": 0.06477675586938858, + "learning_rate": 0.000398, + "loss": 0.0197, + "step": 7960 + }, + { + "epoch": 11.984962406015038, + "grad_norm": 0.08439268916845322, + "learning_rate": 0.00039850000000000004, + "loss": 0.0166, + "step": 7970 + }, + { + "epoch": 12.0, + "grad_norm": 0.05665082857012749, + "learning_rate": 0.00039900000000000005, + "loss": 0.0166, + "step": 7980 + }, + { + "epoch": 12.015037593984962, + "grad_norm": 0.058893561363220215, + "learning_rate": 0.0003995, + "loss": 0.0132, + "step": 7990 + }, + { + "epoch": 12.030075187969924, + "grad_norm": 0.04399947449564934, + "learning_rate": 0.0004, + "loss": 0.0155, + "step": 8000 + }, + { + "epoch": 12.030075187969924, + "eval_cer": 0.02051645791368521, + "eval_loss": 0.058879729360342026, + "eval_runtime": 160.7794, + "eval_samples_per_second": 99.683, + "eval_steps_per_second": 0.784, + "eval_wer": 0.06809938168354977, + "step": 8000 + }, + { + "epoch": 12.045112781954888, + "grad_norm": 0.07498139142990112, + "learning_rate": 0.00040050000000000003, + "loss": 0.0141, + "step": 8010 + }, + { + "epoch": 12.06015037593985, + "grad_norm": 0.08380801230669022, + "learning_rate": 0.00040100000000000004, + "loss": 0.013, + "step": 8020 + }, + { + "epoch": 12.075187969924812, + "grad_norm": 0.08174639195203781, + "learning_rate": 0.00040150000000000006, + "loss": 0.0142, + "step": 8030 + }, + { + "epoch": 12.090225563909774, + "grad_norm": 0.0528586246073246, + "learning_rate": 0.000402, + "loss": 0.0158, + "step": 8040 + }, + { + "epoch": 12.105263157894736, + "grad_norm": 0.05917321890592575, + "learning_rate": 0.0004025, + "loss": 0.0154, + "step": 8050 + }, + { + "epoch": 12.1203007518797, + "grad_norm": 0.044992972165346146, + "learning_rate": 0.00040300000000000004, + "loss": 0.0152, + "step": 8060 + }, + { + "epoch": 12.135338345864662, + "grad_norm": 0.06975208222866058, + "learning_rate": 0.00040350000000000005, + "loss": 0.0157, + "step": 8070 + }, + { + "epoch": 12.150375939849624, + "grad_norm": 0.05955854803323746, + "learning_rate": 0.000404, + "loss": 0.0147, + "step": 8080 + }, + { + "epoch": 12.165413533834586, + "grad_norm": 0.04892636463046074, + "learning_rate": 0.0004045, + "loss": 0.0147, + "step": 8090 + }, + { + "epoch": 12.180451127819548, + "grad_norm": 0.05483972653746605, + "learning_rate": 0.00040500000000000003, + "loss": 0.0164, + "step": 8100 + }, + { + "epoch": 12.19548872180451, + "grad_norm": 0.046551283448934555, + "learning_rate": 0.00040550000000000004, + "loss": 0.0137, + "step": 8110 + }, + { + "epoch": 12.210526315789474, + "grad_norm": 0.05125221610069275, + "learning_rate": 0.00040600000000000006, + "loss": 0.0158, + "step": 8120 + }, + { + "epoch": 12.225563909774436, + "grad_norm": 0.04963212460279465, + "learning_rate": 0.00040649999999999996, + "loss": 0.0126, + "step": 8130 + }, + { + "epoch": 12.240601503759398, + "grad_norm": 0.07371821999549866, + "learning_rate": 0.00040699999999999997, + "loss": 0.0148, + "step": 8140 + }, + { + "epoch": 12.25563909774436, + "grad_norm": 0.0632639080286026, + "learning_rate": 0.0004075, + "loss": 0.0154, + "step": 8150 + }, + { + "epoch": 12.270676691729323, + "grad_norm": 0.07732667773962021, + "learning_rate": 0.000408, + "loss": 0.0153, + "step": 8160 + }, + { + "epoch": 12.285714285714286, + "grad_norm": 0.05606967583298683, + "learning_rate": 0.0004085, + "loss": 0.0145, + "step": 8170 + }, + { + "epoch": 12.300751879699249, + "grad_norm": 0.06270197778940201, + "learning_rate": 0.00040899999999999997, + "loss": 0.0166, + "step": 8180 + }, + { + "epoch": 12.31578947368421, + "grad_norm": 0.06792162358760834, + "learning_rate": 0.0004095, + "loss": 0.0147, + "step": 8190 + }, + { + "epoch": 12.330827067669173, + "grad_norm": 0.06799659878015518, + "learning_rate": 0.00041, + "loss": 0.0154, + "step": 8200 + }, + { + "epoch": 12.345864661654135, + "grad_norm": 0.056890543550252914, + "learning_rate": 0.0004105, + "loss": 0.016, + "step": 8210 + }, + { + "epoch": 12.360902255639097, + "grad_norm": 0.0609557144343853, + "learning_rate": 0.00041099999999999996, + "loss": 0.0159, + "step": 8220 + }, + { + "epoch": 12.37593984962406, + "grad_norm": 0.04662119597196579, + "learning_rate": 0.0004115, + "loss": 0.0132, + "step": 8230 + }, + { + "epoch": 12.390977443609023, + "grad_norm": 0.08694552630186081, + "learning_rate": 0.000412, + "loss": 0.0155, + "step": 8240 + }, + { + "epoch": 12.406015037593985, + "grad_norm": 0.06744252890348434, + "learning_rate": 0.0004125, + "loss": 0.0146, + "step": 8250 + }, + { + "epoch": 12.421052631578947, + "grad_norm": 0.05434747040271759, + "learning_rate": 0.000413, + "loss": 0.0181, + "step": 8260 + }, + { + "epoch": 12.436090225563909, + "grad_norm": 0.05837763100862503, + "learning_rate": 0.00041349999999999997, + "loss": 0.0145, + "step": 8270 + }, + { + "epoch": 12.451127819548873, + "grad_norm": 0.056206777691841125, + "learning_rate": 0.000414, + "loss": 0.015, + "step": 8280 + }, + { + "epoch": 12.466165413533835, + "grad_norm": 0.07385284453630447, + "learning_rate": 0.0004145, + "loss": 0.0161, + "step": 8290 + }, + { + "epoch": 12.481203007518797, + "grad_norm": 0.06202505901455879, + "learning_rate": 0.000415, + "loss": 0.0162, + "step": 8300 + }, + { + "epoch": 12.496240601503759, + "grad_norm": 0.06885461509227753, + "learning_rate": 0.00041549999999999996, + "loss": 0.0179, + "step": 8310 + }, + { + "epoch": 12.511278195488721, + "grad_norm": 0.0704217329621315, + "learning_rate": 0.000416, + "loss": 0.0138, + "step": 8320 + }, + { + "epoch": 12.526315789473685, + "grad_norm": 0.0672963410615921, + "learning_rate": 0.0004165, + "loss": 0.0178, + "step": 8330 + }, + { + "epoch": 12.541353383458647, + "grad_norm": 0.04962855949997902, + "learning_rate": 0.000417, + "loss": 0.0144, + "step": 8340 + }, + { + "epoch": 12.556390977443609, + "grad_norm": 0.07197018712759018, + "learning_rate": 0.0004175, + "loss": 0.015, + "step": 8350 + }, + { + "epoch": 12.571428571428571, + "grad_norm": 0.05259184166789055, + "learning_rate": 0.00041799999999999997, + "loss": 0.015, + "step": 8360 + }, + { + "epoch": 12.586466165413533, + "grad_norm": 0.09718608856201172, + "learning_rate": 0.0004185, + "loss": 0.0154, + "step": 8370 + }, + { + "epoch": 12.601503759398497, + "grad_norm": 0.049638532102108, + "learning_rate": 0.000419, + "loss": 0.0147, + "step": 8380 + }, + { + "epoch": 12.61654135338346, + "grad_norm": 0.06316649168729782, + "learning_rate": 0.0004195, + "loss": 0.0162, + "step": 8390 + }, + { + "epoch": 12.631578947368421, + "grad_norm": 0.06570911407470703, + "learning_rate": 0.00042, + "loss": 0.0126, + "step": 8400 + }, + { + "epoch": 12.646616541353383, + "grad_norm": 0.05419382452964783, + "learning_rate": 0.0004205, + "loss": 0.0163, + "step": 8410 + }, + { + "epoch": 12.661654135338345, + "grad_norm": 0.053494829684495926, + "learning_rate": 0.000421, + "loss": 0.0172, + "step": 8420 + }, + { + "epoch": 12.676691729323307, + "grad_norm": 0.07974492758512497, + "learning_rate": 0.0004215, + "loss": 0.0159, + "step": 8430 + }, + { + "epoch": 12.691729323308271, + "grad_norm": 0.08963408321142197, + "learning_rate": 0.000422, + "loss": 0.0126, + "step": 8440 + }, + { + "epoch": 12.706766917293233, + "grad_norm": 0.042577583342790604, + "learning_rate": 0.00042249999999999997, + "loss": 0.0149, + "step": 8450 + }, + { + "epoch": 12.721804511278195, + "grad_norm": 0.06286728382110596, + "learning_rate": 0.000423, + "loss": 0.0166, + "step": 8460 + }, + { + "epoch": 12.736842105263158, + "grad_norm": 0.05724182724952698, + "learning_rate": 0.0004235, + "loss": 0.0151, + "step": 8470 + }, + { + "epoch": 12.75187969924812, + "grad_norm": 0.09984349459409714, + "learning_rate": 0.000424, + "loss": 0.0161, + "step": 8480 + }, + { + "epoch": 12.766917293233083, + "grad_norm": 0.05217945948243141, + "learning_rate": 0.0004245, + "loss": 0.0164, + "step": 8490 + }, + { + "epoch": 12.781954887218046, + "grad_norm": 0.1186133325099945, + "learning_rate": 0.000425, + "loss": 0.0188, + "step": 8500 + }, + { + "epoch": 12.796992481203008, + "grad_norm": 0.0746011957526207, + "learning_rate": 0.0004255, + "loss": 0.0154, + "step": 8510 + }, + { + "epoch": 12.81203007518797, + "grad_norm": 0.06416549533605576, + "learning_rate": 0.000426, + "loss": 0.0158, + "step": 8520 + }, + { + "epoch": 12.827067669172932, + "grad_norm": 0.07415299862623215, + "learning_rate": 0.0004265, + "loss": 0.0157, + "step": 8530 + }, + { + "epoch": 12.842105263157894, + "grad_norm": 0.059171222150325775, + "learning_rate": 0.000427, + "loss": 0.0146, + "step": 8540 + }, + { + "epoch": 12.857142857142858, + "grad_norm": 0.048288989812135696, + "learning_rate": 0.0004275, + "loss": 0.0145, + "step": 8550 + }, + { + "epoch": 12.87218045112782, + "grad_norm": 0.04808344319462776, + "learning_rate": 0.000428, + "loss": 0.0148, + "step": 8560 + }, + { + "epoch": 12.887218045112782, + "grad_norm": 0.05173613503575325, + "learning_rate": 0.0004285, + "loss": 0.0155, + "step": 8570 + }, + { + "epoch": 12.902255639097744, + "grad_norm": 0.05433941259980202, + "learning_rate": 0.000429, + "loss": 0.0161, + "step": 8580 + }, + { + "epoch": 12.917293233082706, + "grad_norm": 0.06362587958574295, + "learning_rate": 0.0004295, + "loss": 0.016, + "step": 8590 + }, + { + "epoch": 12.93233082706767, + "grad_norm": 0.0658130794763565, + "learning_rate": 0.00043, + "loss": 0.0167, + "step": 8600 + }, + { + "epoch": 12.947368421052632, + "grad_norm": 0.04550983011722565, + "learning_rate": 0.0004305, + "loss": 0.0155, + "step": 8610 + }, + { + "epoch": 12.962406015037594, + "grad_norm": 0.08119028061628342, + "learning_rate": 0.000431, + "loss": 0.0144, + "step": 8620 + }, + { + "epoch": 12.977443609022556, + "grad_norm": 0.058119840919971466, + "learning_rate": 0.0004315, + "loss": 0.0134, + "step": 8630 + }, + { + "epoch": 12.992481203007518, + "grad_norm": 0.06672415882349014, + "learning_rate": 0.000432, + "loss": 0.0148, + "step": 8640 + }, + { + "epoch": 13.007518796992482, + "grad_norm": 0.08687225729227066, + "learning_rate": 0.0004325, + "loss": 0.0156, + "step": 8650 + }, + { + "epoch": 13.022556390977444, + "grad_norm": 0.05837656930088997, + "learning_rate": 0.000433, + "loss": 0.012, + "step": 8660 + }, + { + "epoch": 13.037593984962406, + "grad_norm": 0.06473557651042938, + "learning_rate": 0.0004335, + "loss": 0.012, + "step": 8670 + }, + { + "epoch": 13.052631578947368, + "grad_norm": 0.04419999569654465, + "learning_rate": 0.00043400000000000003, + "loss": 0.0117, + "step": 8680 + }, + { + "epoch": 13.06766917293233, + "grad_norm": 0.06705644726753235, + "learning_rate": 0.0004345, + "loss": 0.0113, + "step": 8690 + }, + { + "epoch": 13.082706766917294, + "grad_norm": 0.057164911180734634, + "learning_rate": 0.000435, + "loss": 0.0109, + "step": 8700 + }, + { + "epoch": 13.097744360902256, + "grad_norm": 0.07219601422548294, + "learning_rate": 0.0004355, + "loss": 0.0136, + "step": 8710 + }, + { + "epoch": 13.112781954887218, + "grad_norm": 0.060701221227645874, + "learning_rate": 0.000436, + "loss": 0.0116, + "step": 8720 + }, + { + "epoch": 13.12781954887218, + "grad_norm": 0.058637991547584534, + "learning_rate": 0.0004365, + "loss": 0.0117, + "step": 8730 + }, + { + "epoch": 13.142857142857142, + "grad_norm": 0.08248470723628998, + "learning_rate": 0.000437, + "loss": 0.0103, + "step": 8740 + }, + { + "epoch": 13.157894736842104, + "grad_norm": 0.11128250509500504, + "learning_rate": 0.0004375, + "loss": 0.0124, + "step": 8750 + }, + { + "epoch": 13.172932330827068, + "grad_norm": 0.06819871068000793, + "learning_rate": 0.000438, + "loss": 0.0128, + "step": 8760 + }, + { + "epoch": 13.18796992481203, + "grad_norm": 0.04775603115558624, + "learning_rate": 0.00043850000000000003, + "loss": 0.0114, + "step": 8770 + }, + { + "epoch": 13.203007518796992, + "grad_norm": 0.06324049830436707, + "learning_rate": 0.000439, + "loss": 0.013, + "step": 8780 + }, + { + "epoch": 13.218045112781954, + "grad_norm": 0.07742930203676224, + "learning_rate": 0.0004395, + "loss": 0.0132, + "step": 8790 + }, + { + "epoch": 13.233082706766917, + "grad_norm": 0.06062326207756996, + "learning_rate": 0.00044, + "loss": 0.0133, + "step": 8800 + }, + { + "epoch": 13.24812030075188, + "grad_norm": 0.0642896518111229, + "learning_rate": 0.00044050000000000003, + "loss": 0.0121, + "step": 8810 + }, + { + "epoch": 13.263157894736842, + "grad_norm": 0.059220731258392334, + "learning_rate": 0.000441, + "loss": 0.0112, + "step": 8820 + }, + { + "epoch": 13.278195488721805, + "grad_norm": 0.0711437240242958, + "learning_rate": 0.0004415, + "loss": 0.0134, + "step": 8830 + }, + { + "epoch": 13.293233082706767, + "grad_norm": 0.055094484239816666, + "learning_rate": 0.000442, + "loss": 0.0123, + "step": 8840 + }, + { + "epoch": 13.308270676691729, + "grad_norm": 0.05611947923898697, + "learning_rate": 0.0004425, + "loss": 0.0123, + "step": 8850 + }, + { + "epoch": 13.323308270676693, + "grad_norm": 0.07001008093357086, + "learning_rate": 0.00044300000000000003, + "loss": 0.012, + "step": 8860 + }, + { + "epoch": 13.338345864661655, + "grad_norm": 0.054901983588933945, + "learning_rate": 0.0004435, + "loss": 0.0132, + "step": 8870 + }, + { + "epoch": 13.353383458646617, + "grad_norm": 0.08323562145233154, + "learning_rate": 0.000444, + "loss": 0.012, + "step": 8880 + }, + { + "epoch": 13.368421052631579, + "grad_norm": 0.07770878076553345, + "learning_rate": 0.0004445, + "loss": 0.0128, + "step": 8890 + }, + { + "epoch": 13.38345864661654, + "grad_norm": 0.06727772951126099, + "learning_rate": 0.00044500000000000003, + "loss": 0.0105, + "step": 8900 + }, + { + "epoch": 13.398496240601503, + "grad_norm": 0.080037921667099, + "learning_rate": 0.00044550000000000004, + "loss": 0.0144, + "step": 8910 + }, + { + "epoch": 13.413533834586467, + "grad_norm": 0.0699772983789444, + "learning_rate": 0.000446, + "loss": 0.0146, + "step": 8920 + }, + { + "epoch": 13.428571428571429, + "grad_norm": 0.06747730076313019, + "learning_rate": 0.0004465, + "loss": 0.013, + "step": 8930 + }, + { + "epoch": 13.443609022556391, + "grad_norm": 0.09341194480657578, + "learning_rate": 0.000447, + "loss": 0.0117, + "step": 8940 + }, + { + "epoch": 13.458646616541353, + "grad_norm": 0.05952294543385506, + "learning_rate": 0.00044750000000000004, + "loss": 0.0146, + "step": 8950 + }, + { + "epoch": 13.473684210526315, + "grad_norm": 0.0465039387345314, + "learning_rate": 0.000448, + "loss": 0.0137, + "step": 8960 + }, + { + "epoch": 13.488721804511279, + "grad_norm": 0.04727974161505699, + "learning_rate": 0.0004485, + "loss": 0.0116, + "step": 8970 + }, + { + "epoch": 13.503759398496241, + "grad_norm": 0.04726318642497063, + "learning_rate": 0.000449, + "loss": 0.015, + "step": 8980 + }, + { + "epoch": 13.518796992481203, + "grad_norm": 0.045061398297548294, + "learning_rate": 0.00044950000000000003, + "loss": 0.013, + "step": 8990 + }, + { + "epoch": 13.533834586466165, + "grad_norm": 0.06756097823381424, + "learning_rate": 0.00045000000000000004, + "loss": 0.0143, + "step": 9000 + }, + { + "epoch": 13.533834586466165, + "eval_cer": 0.022350555397632605, + "eval_loss": 0.06154884770512581, + "eval_runtime": 162.7373, + "eval_samples_per_second": 98.484, + "eval_steps_per_second": 0.774, + "eval_wer": 0.0727273581536958, + "step": 9000 + }, + { + "epoch": 13.548872180451127, + "grad_norm": 0.053801801055669785, + "learning_rate": 0.0004505, + "loss": 0.0136, + "step": 9010 + }, + { + "epoch": 13.563909774436091, + "grad_norm": 0.04228867217898369, + "learning_rate": 0.000451, + "loss": 0.0127, + "step": 9020 + }, + { + "epoch": 13.578947368421053, + "grad_norm": 0.04548037424683571, + "learning_rate": 0.0004515, + "loss": 0.0133, + "step": 9030 + }, + { + "epoch": 13.593984962406015, + "grad_norm": 0.07253804802894592, + "learning_rate": 0.00045200000000000004, + "loss": 0.014, + "step": 9040 + }, + { + "epoch": 13.609022556390977, + "grad_norm": 0.05819770321249962, + "learning_rate": 0.00045250000000000005, + "loss": 0.0147, + "step": 9050 + }, + { + "epoch": 13.62406015037594, + "grad_norm": 0.07398857921361923, + "learning_rate": 0.000453, + "loss": 0.013, + "step": 9060 + }, + { + "epoch": 13.639097744360903, + "grad_norm": 0.06199869140982628, + "learning_rate": 0.0004535, + "loss": 0.0118, + "step": 9070 + }, + { + "epoch": 13.654135338345865, + "grad_norm": 0.06446758657693863, + "learning_rate": 0.00045400000000000003, + "loss": 0.0145, + "step": 9080 + }, + { + "epoch": 13.669172932330827, + "grad_norm": 0.05262603238224983, + "learning_rate": 0.00045450000000000004, + "loss": 0.0151, + "step": 9090 + }, + { + "epoch": 13.68421052631579, + "grad_norm": 0.059041112661361694, + "learning_rate": 0.000455, + "loss": 0.0157, + "step": 9100 + }, + { + "epoch": 13.699248120300751, + "grad_norm": 0.06567554920911789, + "learning_rate": 0.0004555, + "loss": 0.0122, + "step": 9110 + }, + { + "epoch": 13.714285714285714, + "grad_norm": 0.07322700321674347, + "learning_rate": 0.000456, + "loss": 0.0143, + "step": 9120 + }, + { + "epoch": 13.729323308270677, + "grad_norm": 0.045058514922857285, + "learning_rate": 0.00045650000000000004, + "loss": 0.0144, + "step": 9130 + }, + { + "epoch": 13.74436090225564, + "grad_norm": 0.04368934407830238, + "learning_rate": 0.00045700000000000005, + "loss": 0.0112, + "step": 9140 + }, + { + "epoch": 13.759398496240602, + "grad_norm": 0.05394401773810387, + "learning_rate": 0.0004575, + "loss": 0.0139, + "step": 9150 + }, + { + "epoch": 13.774436090225564, + "grad_norm": 0.07156209647655487, + "learning_rate": 0.000458, + "loss": 0.0144, + "step": 9160 + }, + { + "epoch": 13.789473684210526, + "grad_norm": 0.05027615278959274, + "learning_rate": 0.00045850000000000003, + "loss": 0.013, + "step": 9170 + }, + { + "epoch": 13.80451127819549, + "grad_norm": 0.09548698365688324, + "learning_rate": 0.00045900000000000004, + "loss": 0.0146, + "step": 9180 + }, + { + "epoch": 13.819548872180452, + "grad_norm": 0.04820404201745987, + "learning_rate": 0.00045950000000000006, + "loss": 0.0117, + "step": 9190 + }, + { + "epoch": 13.834586466165414, + "grad_norm": 0.05904980003833771, + "learning_rate": 0.00046, + "loss": 0.0146, + "step": 9200 + }, + { + "epoch": 13.849624060150376, + "grad_norm": 0.07122162729501724, + "learning_rate": 0.0004605, + "loss": 0.0151, + "step": 9210 + }, + { + "epoch": 13.864661654135338, + "grad_norm": 0.04693657159805298, + "learning_rate": 0.00046100000000000004, + "loss": 0.0145, + "step": 9220 + }, + { + "epoch": 13.8796992481203, + "grad_norm": 0.09083355963230133, + "learning_rate": 0.00046150000000000005, + "loss": 0.0159, + "step": 9230 + }, + { + "epoch": 13.894736842105264, + "grad_norm": 0.07829580456018448, + "learning_rate": 0.000462, + "loss": 0.0139, + "step": 9240 + }, + { + "epoch": 13.909774436090226, + "grad_norm": 0.0682314783334732, + "learning_rate": 0.0004625, + "loss": 0.013, + "step": 9250 + }, + { + "epoch": 13.924812030075188, + "grad_norm": 0.06102532520890236, + "learning_rate": 0.00046300000000000003, + "loss": 0.0125, + "step": 9260 + }, + { + "epoch": 13.93984962406015, + "grad_norm": 0.05448983982205391, + "learning_rate": 0.00046350000000000004, + "loss": 0.0138, + "step": 9270 + }, + { + "epoch": 13.954887218045112, + "grad_norm": 0.07198873907327652, + "learning_rate": 0.00046400000000000006, + "loss": 0.0138, + "step": 9280 + }, + { + "epoch": 13.969924812030076, + "grad_norm": 0.05566171929240227, + "learning_rate": 0.0004645, + "loss": 0.0129, + "step": 9290 + }, + { + "epoch": 13.984962406015038, + "grad_norm": 0.04728633537888527, + "learning_rate": 0.000465, + "loss": 0.0131, + "step": 9300 + }, + { + "epoch": 14.0, + "grad_norm": 0.08018644154071808, + "learning_rate": 0.00046550000000000004, + "loss": 0.0168, + "step": 9310 + }, + { + "epoch": 14.015037593984962, + "grad_norm": 0.0569804348051548, + "learning_rate": 0.00046600000000000005, + "loss": 0.0105, + "step": 9320 + }, + { + "epoch": 14.030075187969924, + "grad_norm": 0.062460124492645264, + "learning_rate": 0.0004665, + "loss": 0.0109, + "step": 9330 + }, + { + "epoch": 14.045112781954888, + "grad_norm": 0.051482945680618286, + "learning_rate": 0.000467, + "loss": 0.0098, + "step": 9340 + }, + { + "epoch": 14.06015037593985, + "grad_norm": 0.05272672697901726, + "learning_rate": 0.00046750000000000003, + "loss": 0.0095, + "step": 9350 + }, + { + "epoch": 14.075187969924812, + "grad_norm": 0.05545605346560478, + "learning_rate": 0.00046800000000000005, + "loss": 0.0108, + "step": 9360 + }, + { + "epoch": 14.090225563909774, + "grad_norm": 0.06198912113904953, + "learning_rate": 0.00046850000000000006, + "loss": 0.0095, + "step": 9370 + }, + { + "epoch": 14.105263157894736, + "grad_norm": 0.04410150274634361, + "learning_rate": 0.00046899999999999996, + "loss": 0.0097, + "step": 9380 + }, + { + "epoch": 14.1203007518797, + "grad_norm": 0.05411685258150101, + "learning_rate": 0.0004695, + "loss": 0.0105, + "step": 9390 + }, + { + "epoch": 14.135338345864662, + "grad_norm": 0.07978096604347229, + "learning_rate": 0.00047, + "loss": 0.0099, + "step": 9400 + }, + { + "epoch": 14.150375939849624, + "grad_norm": 0.04657706990838051, + "learning_rate": 0.0004705, + "loss": 0.0111, + "step": 9410 + }, + { + "epoch": 14.165413533834586, + "grad_norm": 0.09352941066026688, + "learning_rate": 0.000471, + "loss": 0.0113, + "step": 9420 + }, + { + "epoch": 14.180451127819548, + "grad_norm": 0.049501314759254456, + "learning_rate": 0.00047149999999999997, + "loss": 0.0116, + "step": 9430 + }, + { + "epoch": 14.19548872180451, + "grad_norm": 0.07065962255001068, + "learning_rate": 0.000472, + "loss": 0.0112, + "step": 9440 + }, + { + "epoch": 14.210526315789474, + "grad_norm": 0.04040437564253807, + "learning_rate": 0.0004725, + "loss": 0.0113, + "step": 9450 + }, + { + "epoch": 14.225563909774436, + "grad_norm": 0.0538034662604332, + "learning_rate": 0.000473, + "loss": 0.0101, + "step": 9460 + }, + { + "epoch": 14.240601503759398, + "grad_norm": 0.07104087620973587, + "learning_rate": 0.00047349999999999996, + "loss": 0.0097, + "step": 9470 + }, + { + "epoch": 14.25563909774436, + "grad_norm": 0.05569196119904518, + "learning_rate": 0.000474, + "loss": 0.0107, + "step": 9480 + }, + { + "epoch": 14.270676691729323, + "grad_norm": 0.07058149576187134, + "learning_rate": 0.0004745, + "loss": 0.011, + "step": 9490 + }, + { + "epoch": 14.285714285714286, + "grad_norm": 0.042804960161447525, + "learning_rate": 0.000475, + "loss": 0.0098, + "step": 9500 + }, + { + "epoch": 14.300751879699249, + "grad_norm": 0.05729738995432854, + "learning_rate": 0.0004755, + "loss": 0.0093, + "step": 9510 + }, + { + "epoch": 14.31578947368421, + "grad_norm": 0.06033680588006973, + "learning_rate": 0.00047599999999999997, + "loss": 0.0103, + "step": 9520 + }, + { + "epoch": 14.330827067669173, + "grad_norm": 0.06451141834259033, + "learning_rate": 0.0004765, + "loss": 0.0128, + "step": 9530 + }, + { + "epoch": 14.345864661654135, + "grad_norm": 0.05310095474123955, + "learning_rate": 0.000477, + "loss": 0.0144, + "step": 9540 + }, + { + "epoch": 14.360902255639097, + "grad_norm": 0.055733900517225266, + "learning_rate": 0.0004775, + "loss": 0.0123, + "step": 9550 + }, + { + "epoch": 14.37593984962406, + "grad_norm": 0.04527348652482033, + "learning_rate": 0.00047799999999999996, + "loss": 0.0096, + "step": 9560 + }, + { + "epoch": 14.390977443609023, + "grad_norm": 0.04905983433127403, + "learning_rate": 0.0004785, + "loss": 0.0102, + "step": 9570 + }, + { + "epoch": 14.406015037593985, + "grad_norm": 0.07908350974321365, + "learning_rate": 0.000479, + "loss": 0.0126, + "step": 9580 + }, + { + "epoch": 14.421052631578947, + "grad_norm": 0.07588250190019608, + "learning_rate": 0.0004795, + "loss": 0.0118, + "step": 9590 + }, + { + "epoch": 14.436090225563909, + "grad_norm": 0.045016732066869736, + "learning_rate": 0.00048, + "loss": 0.0112, + "step": 9600 + }, + { + "epoch": 14.451127819548873, + "grad_norm": 0.05452529340982437, + "learning_rate": 0.00048049999999999997, + "loss": 0.0111, + "step": 9610 + }, + { + "epoch": 14.466165413533835, + "grad_norm": 0.05283180996775627, + "learning_rate": 0.000481, + "loss": 0.0112, + "step": 9620 + }, + { + "epoch": 14.481203007518797, + "grad_norm": 0.07583435624837875, + "learning_rate": 0.0004815, + "loss": 0.0121, + "step": 9630 + }, + { + "epoch": 14.496240601503759, + "grad_norm": 0.057554978877305984, + "learning_rate": 0.000482, + "loss": 0.0119, + "step": 9640 + }, + { + "epoch": 14.511278195488721, + "grad_norm": 0.055423881858587265, + "learning_rate": 0.0004825, + "loss": 0.0111, + "step": 9650 + }, + { + "epoch": 14.526315789473685, + "grad_norm": 0.057395271956920624, + "learning_rate": 0.000483, + "loss": 0.0116, + "step": 9660 + }, + { + "epoch": 14.541353383458647, + "grad_norm": 0.07486916333436966, + "learning_rate": 0.0004835, + "loss": 0.0129, + "step": 9670 + }, + { + "epoch": 14.556390977443609, + "grad_norm": 0.0678727999329567, + "learning_rate": 0.000484, + "loss": 0.0121, + "step": 9680 + }, + { + "epoch": 14.571428571428571, + "grad_norm": 0.04931911826133728, + "learning_rate": 0.0004845, + "loss": 0.0115, + "step": 9690 + }, + { + "epoch": 14.586466165413533, + "grad_norm": 0.100761778652668, + "learning_rate": 0.00048499999999999997, + "loss": 0.0108, + "step": 9700 + }, + { + "epoch": 14.601503759398497, + "grad_norm": 0.08065155893564224, + "learning_rate": 0.0004855, + "loss": 0.01, + "step": 9710 + }, + { + "epoch": 14.61654135338346, + "grad_norm": 0.07073182612657547, + "learning_rate": 0.000486, + "loss": 0.0115, + "step": 9720 + }, + { + "epoch": 14.631578947368421, + "grad_norm": 0.06363115459680557, + "learning_rate": 0.0004865, + "loss": 0.013, + "step": 9730 + }, + { + "epoch": 14.646616541353383, + "grad_norm": 0.047148413956165314, + "learning_rate": 0.000487, + "loss": 0.011, + "step": 9740 + }, + { + "epoch": 14.661654135338345, + "grad_norm": 0.057262104004621506, + "learning_rate": 0.0004875, + "loss": 0.013, + "step": 9750 + }, + { + "epoch": 14.676691729323307, + "grad_norm": 0.09763623028993607, + "learning_rate": 0.000488, + "loss": 0.0115, + "step": 9760 + }, + { + "epoch": 14.691729323308271, + "grad_norm": 0.09714463353157043, + "learning_rate": 0.0004885, + "loss": 0.0117, + "step": 9770 + }, + { + "epoch": 14.706766917293233, + "grad_norm": 0.0539412647485733, + "learning_rate": 0.000489, + "loss": 0.0119, + "step": 9780 + }, + { + "epoch": 14.721804511278195, + "grad_norm": 0.05679219961166382, + "learning_rate": 0.0004895, + "loss": 0.0114, + "step": 9790 + }, + { + "epoch": 14.736842105263158, + "grad_norm": 0.04313800483942032, + "learning_rate": 0.00049, + "loss": 0.0117, + "step": 9800 + }, + { + "epoch": 14.75187969924812, + "grad_norm": 0.04520241171121597, + "learning_rate": 0.0004905, + "loss": 0.0119, + "step": 9810 + }, + { + "epoch": 14.766917293233083, + "grad_norm": 0.08428625762462616, + "learning_rate": 0.000491, + "loss": 0.0107, + "step": 9820 + }, + { + "epoch": 14.781954887218046, + "grad_norm": 0.05508558824658394, + "learning_rate": 0.0004915, + "loss": 0.0143, + "step": 9830 + }, + { + "epoch": 14.796992481203008, + "grad_norm": 0.06821640580892563, + "learning_rate": 0.000492, + "loss": 0.0117, + "step": 9840 + }, + { + "epoch": 14.81203007518797, + "grad_norm": 0.05724147707223892, + "learning_rate": 0.0004925, + "loss": 0.0104, + "step": 9850 + }, + { + "epoch": 14.827067669172932, + "grad_norm": 0.0468720905482769, + "learning_rate": 0.0004930000000000001, + "loss": 0.0114, + "step": 9860 + }, + { + "epoch": 14.842105263157894, + "grad_norm": 0.04705962911248207, + "learning_rate": 0.0004935, + "loss": 0.0116, + "step": 9870 + }, + { + "epoch": 14.857142857142858, + "grad_norm": 0.04345636069774628, + "learning_rate": 0.000494, + "loss": 0.011, + "step": 9880 + }, + { + "epoch": 14.87218045112782, + "grad_norm": 0.06471360474824905, + "learning_rate": 0.0004945, + "loss": 0.0124, + "step": 9890 + }, + { + "epoch": 14.887218045112782, + "grad_norm": 0.09895563870668411, + "learning_rate": 0.000495, + "loss": 0.0114, + "step": 9900 + }, + { + "epoch": 14.902255639097744, + "grad_norm": 0.06212034448981285, + "learning_rate": 0.0004955, + "loss": 0.012, + "step": 9910 + }, + { + "epoch": 14.917293233082706, + "grad_norm": 0.03857619687914848, + "learning_rate": 0.000496, + "loss": 0.0123, + "step": 9920 + }, + { + "epoch": 14.93233082706767, + "grad_norm": 0.03977648541331291, + "learning_rate": 0.0004965, + "loss": 0.0132, + "step": 9930 + }, + { + "epoch": 14.947368421052632, + "grad_norm": 0.1021510511636734, + "learning_rate": 0.000497, + "loss": 0.0114, + "step": 9940 + }, + { + "epoch": 14.962406015037594, + "grad_norm": 0.08820920437574387, + "learning_rate": 0.0004975, + "loss": 0.0112, + "step": 9950 + }, + { + "epoch": 14.977443609022556, + "grad_norm": 0.06036331132054329, + "learning_rate": 0.000498, + "loss": 0.013, + "step": 9960 + }, + { + "epoch": 14.992481203007518, + "grad_norm": 0.046254727989435196, + "learning_rate": 0.0004985, + "loss": 0.0104, + "step": 9970 + }, + { + "epoch": 15.007518796992482, + "grad_norm": 0.06697358936071396, + "learning_rate": 0.000499, + "loss": 0.0106, + "step": 9980 + }, + { + "epoch": 15.022556390977444, + "grad_norm": 0.04633962735533714, + "learning_rate": 0.0004995, + "loss": 0.0098, + "step": 9990 + }, + { + "epoch": 15.037593984962406, + "grad_norm": 0.04951399192214012, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 10000 + }, + { + "epoch": 15.037593984962406, + "eval_cer": 0.020932649111377388, + "eval_loss": 0.0679001584649086, + "eval_runtime": 160.9624, + "eval_samples_per_second": 99.57, + "eval_steps_per_second": 0.783, + "eval_wer": 0.07097013663102107, + "step": 10000 + }, + { + "epoch": 15.052631578947368, + "grad_norm": 0.07111406326293945, + "learning_rate": 0.0005, + "loss": 0.0092, + "step": 10010 + }, + { + "epoch": 15.06766917293233, + "grad_norm": 0.05811820924282074, + "learning_rate": 0.0005, + "loss": 0.0082, + "step": 10020 + }, + { + "epoch": 15.082706766917294, + "grad_norm": 0.06517420709133148, + "learning_rate": 0.0005, + "loss": 0.0106, + "step": 10030 + }, + { + "epoch": 15.097744360902256, + "grad_norm": 0.08434344828128815, + "learning_rate": 0.0005, + "loss": 0.0105, + "step": 10040 + }, + { + "epoch": 15.112781954887218, + "grad_norm": 0.05727405101060867, + "learning_rate": 0.0005, + "loss": 0.009, + "step": 10050 + }, + { + "epoch": 15.12781954887218, + "grad_norm": 0.04701713100075722, + "learning_rate": 0.0005, + "loss": 0.0112, + "step": 10060 + }, + { + "epoch": 15.142857142857142, + "grad_norm": 0.07056374102830887, + "learning_rate": 0.0005, + "loss": 0.0109, + "step": 10070 + }, + { + "epoch": 15.157894736842104, + "grad_norm": 0.06422567367553711, + "learning_rate": 0.0005, + "loss": 0.0105, + "step": 10080 + }, + { + "epoch": 15.172932330827068, + "grad_norm": 0.08642735332250595, + "learning_rate": 0.0005, + "loss": 0.0089, + "step": 10090 + }, + { + "epoch": 15.18796992481203, + "grad_norm": 0.055170509964227676, + "learning_rate": 0.0005, + "loss": 0.0102, + "step": 10100 + }, + { + "epoch": 15.203007518796992, + "grad_norm": 0.07998238503932953, + "learning_rate": 0.0005, + "loss": 0.0094, + "step": 10110 + }, + { + "epoch": 15.218045112781954, + "grad_norm": 0.05182688683271408, + "learning_rate": 0.0005, + "loss": 0.0109, + "step": 10120 + }, + { + "epoch": 15.233082706766917, + "grad_norm": 0.06002676114439964, + "learning_rate": 0.0005, + "loss": 0.0103, + "step": 10130 + }, + { + "epoch": 15.24812030075188, + "grad_norm": 0.05442851781845093, + "learning_rate": 0.0005, + "loss": 0.0093, + "step": 10140 + }, + { + "epoch": 15.263157894736842, + "grad_norm": 0.06353688985109329, + "learning_rate": 0.0005, + "loss": 0.0103, + "step": 10150 + }, + { + "epoch": 15.278195488721805, + "grad_norm": 0.04935711994767189, + "learning_rate": 0.0005, + "loss": 0.0082, + "step": 10160 + }, + { + "epoch": 15.293233082706767, + "grad_norm": 0.07050047814846039, + "learning_rate": 0.0005, + "loss": 0.009, + "step": 10170 + }, + { + "epoch": 15.308270676691729, + "grad_norm": 0.10220488160848618, + "learning_rate": 0.0005, + "loss": 0.0096, + "step": 10180 + }, + { + "epoch": 15.323308270676693, + "grad_norm": 0.047242119908332825, + "learning_rate": 0.0005, + "loss": 0.0108, + "step": 10190 + }, + { + "epoch": 15.338345864661655, + "grad_norm": 0.05744297057390213, + "learning_rate": 0.0005, + "loss": 0.0095, + "step": 10200 + }, + { + "epoch": 15.353383458646617, + "grad_norm": 0.05725282058119774, + "learning_rate": 0.0005, + "loss": 0.0104, + "step": 10210 + }, + { + "epoch": 15.368421052631579, + "grad_norm": 0.043894827365875244, + "learning_rate": 0.0005, + "loss": 0.0095, + "step": 10220 + }, + { + "epoch": 15.38345864661654, + "grad_norm": 0.09758011996746063, + "learning_rate": 0.0005, + "loss": 0.0097, + "step": 10230 + }, + { + "epoch": 15.398496240601503, + "grad_norm": 0.07210401445627213, + "learning_rate": 0.0005, + "loss": 0.0098, + "step": 10240 + }, + { + "epoch": 15.413533834586467, + "grad_norm": 0.06916919350624084, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 10250 + }, + { + "epoch": 15.428571428571429, + "grad_norm": 0.05191931873559952, + "learning_rate": 0.0005, + "loss": 0.0097, + "step": 10260 + }, + { + "epoch": 15.443609022556391, + "grad_norm": 0.06755432486534119, + "learning_rate": 0.0005, + "loss": 0.0114, + "step": 10270 + }, + { + "epoch": 15.458646616541353, + "grad_norm": 0.10145001113414764, + "learning_rate": 0.0005, + "loss": 0.0102, + "step": 10280 + }, + { + "epoch": 15.473684210526315, + "grad_norm": 0.09195694327354431, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 10290 + }, + { + "epoch": 15.488721804511279, + "grad_norm": 0.046868231147527695, + "learning_rate": 0.0005, + "loss": 0.0104, + "step": 10300 + }, + { + "epoch": 15.503759398496241, + "grad_norm": 0.06774283200502396, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 10310 + }, + { + "epoch": 15.518796992481203, + "grad_norm": 0.05834305286407471, + "learning_rate": 0.0005, + "loss": 0.0101, + "step": 10320 + }, + { + "epoch": 15.533834586466165, + "grad_norm": 0.05482442304491997, + "learning_rate": 0.0005, + "loss": 0.0089, + "step": 10330 + }, + { + "epoch": 15.548872180451127, + "grad_norm": 0.06710460782051086, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 10340 + }, + { + "epoch": 15.563909774436091, + "grad_norm": 0.12233841419219971, + "learning_rate": 0.0005, + "loss": 0.0095, + "step": 10350 + }, + { + "epoch": 15.578947368421053, + "grad_norm": 0.04426002502441406, + "learning_rate": 0.0005, + "loss": 0.0094, + "step": 10360 + }, + { + "epoch": 15.593984962406015, + "grad_norm": 0.08595938980579376, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 10370 + }, + { + "epoch": 15.609022556390977, + "grad_norm": 0.04955907538533211, + "learning_rate": 0.0005, + "loss": 0.0098, + "step": 10380 + }, + { + "epoch": 15.62406015037594, + "grad_norm": 0.06680439412593842, + "learning_rate": 0.0005, + "loss": 0.011, + "step": 10390 + }, + { + "epoch": 15.639097744360903, + "grad_norm": 0.06889624148607254, + "learning_rate": 0.0005, + "loss": 0.0092, + "step": 10400 + }, + { + "epoch": 15.654135338345865, + "grad_norm": 0.06784632056951523, + "learning_rate": 0.0005, + "loss": 0.0092, + "step": 10410 + }, + { + "epoch": 15.669172932330827, + "grad_norm": 0.05448812246322632, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 10420 + }, + { + "epoch": 15.68421052631579, + "grad_norm": 0.06872930377721786, + "learning_rate": 0.0005, + "loss": 0.0109, + "step": 10430 + }, + { + "epoch": 15.699248120300751, + "grad_norm": 0.04210126772522926, + "learning_rate": 0.0005, + "loss": 0.0087, + "step": 10440 + }, + { + "epoch": 15.714285714285714, + "grad_norm": 0.0591299869120121, + "learning_rate": 0.0005, + "loss": 0.0089, + "step": 10450 + }, + { + "epoch": 15.729323308270677, + "grad_norm": 0.044891711324453354, + "learning_rate": 0.0005, + "loss": 0.0098, + "step": 10460 + }, + { + "epoch": 15.74436090225564, + "grad_norm": 0.07303763926029205, + "learning_rate": 0.0005, + "loss": 0.0091, + "step": 10470 + }, + { + "epoch": 15.759398496240602, + "grad_norm": 0.060911137610673904, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 10480 + }, + { + "epoch": 15.774436090225564, + "grad_norm": 0.05211096629500389, + "learning_rate": 0.0005, + "loss": 0.0105, + "step": 10490 + }, + { + "epoch": 15.789473684210526, + "grad_norm": 0.06013522669672966, + "learning_rate": 0.0005, + "loss": 0.0108, + "step": 10500 + }, + { + "epoch": 15.80451127819549, + "grad_norm": 0.08278252184391022, + "learning_rate": 0.0005, + "loss": 0.0096, + "step": 10510 + }, + { + "epoch": 15.819548872180452, + "grad_norm": 0.06839005649089813, + "learning_rate": 0.0005, + "loss": 0.0097, + "step": 10520 + }, + { + "epoch": 15.834586466165414, + "grad_norm": 0.051177337765693665, + "learning_rate": 0.0005, + "loss": 0.0121, + "step": 10530 + }, + { + "epoch": 15.849624060150376, + "grad_norm": 0.0661792978644371, + "learning_rate": 0.0005, + "loss": 0.0121, + "step": 10540 + }, + { + "epoch": 15.864661654135338, + "grad_norm": 0.05542397499084473, + "learning_rate": 0.0005, + "loss": 0.0098, + "step": 10550 + }, + { + "epoch": 15.8796992481203, + "grad_norm": 0.059120114892721176, + "learning_rate": 0.0005, + "loss": 0.0108, + "step": 10560 + }, + { + "epoch": 15.894736842105264, + "grad_norm": 0.047427501529455185, + "learning_rate": 0.0005, + "loss": 0.0094, + "step": 10570 + }, + { + "epoch": 15.909774436090226, + "grad_norm": 0.04838680103421211, + "learning_rate": 0.0005, + "loss": 0.0091, + "step": 10580 + }, + { + "epoch": 15.924812030075188, + "grad_norm": 0.05245792865753174, + "learning_rate": 0.0005, + "loss": 0.0101, + "step": 10590 + }, + { + "epoch": 15.93984962406015, + "grad_norm": 0.04002607986330986, + "learning_rate": 0.0005, + "loss": 0.0092, + "step": 10600 + }, + { + "epoch": 15.954887218045112, + "grad_norm": 0.0571436807513237, + "learning_rate": 0.0005, + "loss": 0.0093, + "step": 10610 + }, + { + "epoch": 15.969924812030076, + "grad_norm": 0.09747693687677383, + "learning_rate": 0.0005, + "loss": 0.0105, + "step": 10620 + }, + { + "epoch": 15.984962406015038, + "grad_norm": 0.05304805189371109, + "learning_rate": 0.0005, + "loss": 0.0113, + "step": 10630 + }, + { + "epoch": 16.0, + "grad_norm": 0.060404617339372635, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 10640 + }, + { + "epoch": 16.015037593984964, + "grad_norm": 0.04885092377662659, + "learning_rate": 0.0005, + "loss": 0.0082, + "step": 10650 + }, + { + "epoch": 16.030075187969924, + "grad_norm": 0.05729106068611145, + "learning_rate": 0.0005, + "loss": 0.0075, + "step": 10660 + }, + { + "epoch": 16.045112781954888, + "grad_norm": 0.08012731373310089, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 10670 + }, + { + "epoch": 16.06015037593985, + "grad_norm": 0.049254823476076126, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 10680 + }, + { + "epoch": 16.075187969924812, + "grad_norm": 0.044145312160253525, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 10690 + }, + { + "epoch": 16.090225563909776, + "grad_norm": 0.04696081206202507, + "learning_rate": 0.0005, + "loss": 0.0079, + "step": 10700 + }, + { + "epoch": 16.105263157894736, + "grad_norm": 0.045394688844680786, + "learning_rate": 0.0005, + "loss": 0.0062, + "step": 10710 + }, + { + "epoch": 16.1203007518797, + "grad_norm": 0.060016728937625885, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 10720 + }, + { + "epoch": 16.13533834586466, + "grad_norm": 0.041252173483371735, + "learning_rate": 0.0005, + "loss": 0.0082, + "step": 10730 + }, + { + "epoch": 16.150375939849624, + "grad_norm": 0.04355092719197273, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 10740 + }, + { + "epoch": 16.165413533834588, + "grad_norm": 0.05501066520810127, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 10750 + }, + { + "epoch": 16.18045112781955, + "grad_norm": 0.050708189606666565, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 10760 + }, + { + "epoch": 16.195488721804512, + "grad_norm": 0.053117137402296066, + "learning_rate": 0.0005, + "loss": 0.0071, + "step": 10770 + }, + { + "epoch": 16.210526315789473, + "grad_norm": 0.048739053308963776, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 10780 + }, + { + "epoch": 16.225563909774436, + "grad_norm": 0.052768487483263016, + "learning_rate": 0.0005, + "loss": 0.0075, + "step": 10790 + }, + { + "epoch": 16.2406015037594, + "grad_norm": 0.0591270811855793, + "learning_rate": 0.0005, + "loss": 0.0076, + "step": 10800 + }, + { + "epoch": 16.25563909774436, + "grad_norm": 0.050967298448085785, + "learning_rate": 0.0005, + "loss": 0.0084, + "step": 10810 + }, + { + "epoch": 16.270676691729324, + "grad_norm": 0.05149754881858826, + "learning_rate": 0.0005, + "loss": 0.0095, + "step": 10820 + }, + { + "epoch": 16.285714285714285, + "grad_norm": 0.04156196862459183, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 10830 + }, + { + "epoch": 16.30075187969925, + "grad_norm": 0.04769454896450043, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 10840 + }, + { + "epoch": 16.31578947368421, + "grad_norm": 0.06267265975475311, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 10850 + }, + { + "epoch": 16.330827067669173, + "grad_norm": 0.05772874131798744, + "learning_rate": 0.0005, + "loss": 0.0088, + "step": 10860 + }, + { + "epoch": 16.345864661654137, + "grad_norm": 0.05517463758587837, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 10870 + }, + { + "epoch": 16.360902255639097, + "grad_norm": 0.037167418748140335, + "learning_rate": 0.0005, + "loss": 0.0079, + "step": 10880 + }, + { + "epoch": 16.37593984962406, + "grad_norm": 0.06295276433229446, + "learning_rate": 0.0005, + "loss": 0.0088, + "step": 10890 + }, + { + "epoch": 16.39097744360902, + "grad_norm": 0.06420313566923141, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 10900 + }, + { + "epoch": 16.406015037593985, + "grad_norm": 0.04114757850766182, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 10910 + }, + { + "epoch": 16.42105263157895, + "grad_norm": 0.0974854901432991, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 10920 + }, + { + "epoch": 16.43609022556391, + "grad_norm": 0.037163183093070984, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 10930 + }, + { + "epoch": 16.451127819548873, + "grad_norm": 0.04345502331852913, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 10940 + }, + { + "epoch": 16.466165413533833, + "grad_norm": 0.05326547473669052, + "learning_rate": 0.0005, + "loss": 0.009, + "step": 10950 + }, + { + "epoch": 16.481203007518797, + "grad_norm": 0.05854066461324692, + "learning_rate": 0.0005, + "loss": 0.0087, + "step": 10960 + }, + { + "epoch": 16.49624060150376, + "grad_norm": 0.03696458786725998, + "learning_rate": 0.0005, + "loss": 0.0084, + "step": 10970 + }, + { + "epoch": 16.51127819548872, + "grad_norm": 0.08729267865419388, + "learning_rate": 0.0005, + "loss": 0.0091, + "step": 10980 + }, + { + "epoch": 16.526315789473685, + "grad_norm": 0.053089242428541183, + "learning_rate": 0.0005, + "loss": 0.0098, + "step": 10990 + }, + { + "epoch": 16.541353383458645, + "grad_norm": 0.06949713081121445, + "learning_rate": 0.0005, + "loss": 0.0079, + "step": 11000 + }, + { + "epoch": 16.541353383458645, + "eval_cer": 0.019777242502097957, + "eval_loss": 0.07057972252368927, + "eval_runtime": 162.3218, + "eval_samples_per_second": 98.736, + "eval_steps_per_second": 0.776, + "eval_wer": 0.06871487906181285, + "step": 11000 + }, + { + "epoch": 16.55639097744361, + "grad_norm": 0.046286676079034805, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 11010 + }, + { + "epoch": 16.571428571428573, + "grad_norm": 0.05469865724444389, + "learning_rate": 0.0005, + "loss": 0.0093, + "step": 11020 + }, + { + "epoch": 16.586466165413533, + "grad_norm": 0.03843056783080101, + "learning_rate": 0.0005, + "loss": 0.0076, + "step": 11030 + }, + { + "epoch": 16.601503759398497, + "grad_norm": 0.05467670410871506, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11040 + }, + { + "epoch": 16.616541353383457, + "grad_norm": 0.05813300237059593, + "learning_rate": 0.0005, + "loss": 0.009, + "step": 11050 + }, + { + "epoch": 16.63157894736842, + "grad_norm": 0.057580120861530304, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 11060 + }, + { + "epoch": 16.646616541353385, + "grad_norm": 0.0483466237783432, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 11070 + }, + { + "epoch": 16.661654135338345, + "grad_norm": 0.044786520302295685, + "learning_rate": 0.0005, + "loss": 0.009, + "step": 11080 + }, + { + "epoch": 16.67669172932331, + "grad_norm": 0.05274363234639168, + "learning_rate": 0.0005, + "loss": 0.0093, + "step": 11090 + }, + { + "epoch": 16.69172932330827, + "grad_norm": 0.06333556771278381, + "learning_rate": 0.0005, + "loss": 0.0074, + "step": 11100 + }, + { + "epoch": 16.706766917293233, + "grad_norm": 0.05888620391488075, + "learning_rate": 0.0005, + "loss": 0.0103, + "step": 11110 + }, + { + "epoch": 16.721804511278194, + "grad_norm": 0.05396256968379021, + "learning_rate": 0.0005, + "loss": 0.0097, + "step": 11120 + }, + { + "epoch": 16.736842105263158, + "grad_norm": 0.07338481396436691, + "learning_rate": 0.0005, + "loss": 0.0075, + "step": 11130 + }, + { + "epoch": 16.75187969924812, + "grad_norm": 0.046687304973602295, + "learning_rate": 0.0005, + "loss": 0.0084, + "step": 11140 + }, + { + "epoch": 16.76691729323308, + "grad_norm": 0.07993320375680923, + "learning_rate": 0.0005, + "loss": 0.0099, + "step": 11150 + }, + { + "epoch": 16.781954887218046, + "grad_norm": 0.04469626024365425, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 11160 + }, + { + "epoch": 16.796992481203006, + "grad_norm": 0.0511959008872509, + "learning_rate": 0.0005, + "loss": 0.0086, + "step": 11170 + }, + { + "epoch": 16.81203007518797, + "grad_norm": 0.040308043360710144, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 11180 + }, + { + "epoch": 16.827067669172934, + "grad_norm": 0.045106034725904465, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 11190 + }, + { + "epoch": 16.842105263157894, + "grad_norm": 0.06313479691743851, + "learning_rate": 0.0005, + "loss": 0.0094, + "step": 11200 + }, + { + "epoch": 16.857142857142858, + "grad_norm": 0.05153006687760353, + "learning_rate": 0.0005, + "loss": 0.0094, + "step": 11210 + }, + { + "epoch": 16.872180451127818, + "grad_norm": 0.0472828634083271, + "learning_rate": 0.0005, + "loss": 0.0079, + "step": 11220 + }, + { + "epoch": 16.887218045112782, + "grad_norm": 0.04486662521958351, + "learning_rate": 0.0005, + "loss": 0.0097, + "step": 11230 + }, + { + "epoch": 16.902255639097746, + "grad_norm": 0.04471254348754883, + "learning_rate": 0.0005, + "loss": 0.0091, + "step": 11240 + }, + { + "epoch": 16.917293233082706, + "grad_norm": 0.055038612335920334, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 11250 + }, + { + "epoch": 16.93233082706767, + "grad_norm": 0.054290082305669785, + "learning_rate": 0.0005, + "loss": 0.0077, + "step": 11260 + }, + { + "epoch": 16.94736842105263, + "grad_norm": 0.04184836149215698, + "learning_rate": 0.0005, + "loss": 0.0095, + "step": 11270 + }, + { + "epoch": 16.962406015037594, + "grad_norm": 0.06877528131008148, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11280 + }, + { + "epoch": 16.977443609022558, + "grad_norm": 0.04937044531106949, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 11290 + }, + { + "epoch": 16.992481203007518, + "grad_norm": 0.05039536952972412, + "learning_rate": 0.0005, + "loss": 0.0083, + "step": 11300 + }, + { + "epoch": 17.007518796992482, + "grad_norm": 0.08637846261262894, + "learning_rate": 0.0005, + "loss": 0.0089, + "step": 11310 + }, + { + "epoch": 17.022556390977442, + "grad_norm": 0.04532424360513687, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 11320 + }, + { + "epoch": 17.037593984962406, + "grad_norm": 0.06444225460290909, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 11330 + }, + { + "epoch": 17.05263157894737, + "grad_norm": 0.04985102638602257, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 11340 + }, + { + "epoch": 17.06766917293233, + "grad_norm": 0.04609823226928711, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 11350 + }, + { + "epoch": 17.082706766917294, + "grad_norm": 0.03739221766591072, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11360 + }, + { + "epoch": 17.097744360902254, + "grad_norm": 0.045340269804000854, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 11370 + }, + { + "epoch": 17.112781954887218, + "grad_norm": 0.03452746197581291, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 11380 + }, + { + "epoch": 17.127819548872182, + "grad_norm": 0.08674699813127518, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 11390 + }, + { + "epoch": 17.142857142857142, + "grad_norm": 0.03903883323073387, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11400 + }, + { + "epoch": 17.157894736842106, + "grad_norm": 0.060459207743406296, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 11410 + }, + { + "epoch": 17.172932330827066, + "grad_norm": 0.0740484669804573, + "learning_rate": 0.0005, + "loss": 0.0077, + "step": 11420 + }, + { + "epoch": 17.18796992481203, + "grad_norm": 0.05303926020860672, + "learning_rate": 0.0005, + "loss": 0.0066, + "step": 11430 + }, + { + "epoch": 17.203007518796994, + "grad_norm": 0.05927836149930954, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 11440 + }, + { + "epoch": 17.218045112781954, + "grad_norm": 0.03192131593823433, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 11450 + }, + { + "epoch": 17.23308270676692, + "grad_norm": 0.04417894780635834, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 11460 + }, + { + "epoch": 17.24812030075188, + "grad_norm": 0.0631505697965622, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 11470 + }, + { + "epoch": 17.263157894736842, + "grad_norm": 0.05946943163871765, + "learning_rate": 0.0005, + "loss": 0.0074, + "step": 11480 + }, + { + "epoch": 17.278195488721803, + "grad_norm": 0.03662870079278946, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 11490 + }, + { + "epoch": 17.293233082706767, + "grad_norm": 0.045375991612672806, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 11500 + }, + { + "epoch": 17.30827067669173, + "grad_norm": 0.05539242550730705, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 11510 + }, + { + "epoch": 17.32330827067669, + "grad_norm": 0.04769201576709747, + "learning_rate": 0.0005, + "loss": 0.0063, + "step": 11520 + }, + { + "epoch": 17.338345864661655, + "grad_norm": 0.053807903081178665, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 11530 + }, + { + "epoch": 17.353383458646615, + "grad_norm": 0.06600050628185272, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11540 + }, + { + "epoch": 17.36842105263158, + "grad_norm": 0.07080283015966415, + "learning_rate": 0.0005, + "loss": 0.0089, + "step": 11550 + }, + { + "epoch": 17.383458646616543, + "grad_norm": 0.052283599972724915, + "learning_rate": 0.0005, + "loss": 0.0077, + "step": 11560 + }, + { + "epoch": 17.398496240601503, + "grad_norm": 0.05191454291343689, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 11570 + }, + { + "epoch": 17.413533834586467, + "grad_norm": 0.055100537836551666, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 11580 + }, + { + "epoch": 17.428571428571427, + "grad_norm": 0.04330799728631973, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11590 + }, + { + "epoch": 17.44360902255639, + "grad_norm": 0.05173167586326599, + "learning_rate": 0.0005, + "loss": 0.0065, + "step": 11600 + }, + { + "epoch": 17.458646616541355, + "grad_norm": 0.05393906310200691, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 11610 + }, + { + "epoch": 17.473684210526315, + "grad_norm": 0.07504130154848099, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 11620 + }, + { + "epoch": 17.48872180451128, + "grad_norm": 0.05569503828883171, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 11630 + }, + { + "epoch": 17.50375939849624, + "grad_norm": 0.03959733992815018, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 11640 + }, + { + "epoch": 17.518796992481203, + "grad_norm": 0.056042514741420746, + "learning_rate": 0.0005, + "loss": 0.0071, + "step": 11650 + }, + { + "epoch": 17.533834586466167, + "grad_norm": 0.1487104296684265, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11660 + }, + { + "epoch": 17.548872180451127, + "grad_norm": 0.04992859438061714, + "learning_rate": 0.0005, + "loss": 0.0077, + "step": 11670 + }, + { + "epoch": 17.56390977443609, + "grad_norm": 0.04357229173183441, + "learning_rate": 0.0005, + "loss": 0.0076, + "step": 11680 + }, + { + "epoch": 17.57894736842105, + "grad_norm": 0.042481567710638046, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11690 + }, + { + "epoch": 17.593984962406015, + "grad_norm": 0.04641619324684143, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 11700 + }, + { + "epoch": 17.60902255639098, + "grad_norm": 0.06271351128816605, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11710 + }, + { + "epoch": 17.62406015037594, + "grad_norm": 0.1045893132686615, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 11720 + }, + { + "epoch": 17.639097744360903, + "grad_norm": 0.03514842316508293, + "learning_rate": 0.0005, + "loss": 0.0082, + "step": 11730 + }, + { + "epoch": 17.654135338345863, + "grad_norm": 0.05320287123322487, + "learning_rate": 0.0005, + "loss": 0.0079, + "step": 11740 + }, + { + "epoch": 17.669172932330827, + "grad_norm": 0.06244739145040512, + "learning_rate": 0.0005, + "loss": 0.0084, + "step": 11750 + }, + { + "epoch": 17.68421052631579, + "grad_norm": 0.050358764827251434, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 11760 + }, + { + "epoch": 17.69924812030075, + "grad_norm": 0.07427013665437698, + "learning_rate": 0.0005, + "loss": 0.0076, + "step": 11770 + }, + { + "epoch": 17.714285714285715, + "grad_norm": 0.05425640568137169, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 11780 + }, + { + "epoch": 17.729323308270676, + "grad_norm": 0.049843478947877884, + "learning_rate": 0.0005, + "loss": 0.0075, + "step": 11790 + }, + { + "epoch": 17.74436090225564, + "grad_norm": 0.09088059514760971, + "learning_rate": 0.0005, + "loss": 0.0074, + "step": 11800 + }, + { + "epoch": 17.7593984962406, + "grad_norm": 0.05668364837765694, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 11810 + }, + { + "epoch": 17.774436090225564, + "grad_norm": 0.039140280336141586, + "learning_rate": 0.0005, + "loss": 0.0063, + "step": 11820 + }, + { + "epoch": 17.789473684210527, + "grad_norm": 0.049065474420785904, + "learning_rate": 0.0005, + "loss": 0.0076, + "step": 11830 + }, + { + "epoch": 17.804511278195488, + "grad_norm": 0.07399141788482666, + "learning_rate": 0.0005, + "loss": 0.0097, + "step": 11840 + }, + { + "epoch": 17.81954887218045, + "grad_norm": 0.06222929805517197, + "learning_rate": 0.0005, + "loss": 0.0077, + "step": 11850 + }, + { + "epoch": 17.834586466165412, + "grad_norm": 0.06838171929121017, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 11860 + }, + { + "epoch": 17.849624060150376, + "grad_norm": 0.03239224851131439, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 11870 + }, + { + "epoch": 17.86466165413534, + "grad_norm": 0.05101778730750084, + "learning_rate": 0.0005, + "loss": 0.0065, + "step": 11880 + }, + { + "epoch": 17.8796992481203, + "grad_norm": 0.05535724386572838, + "learning_rate": 0.0005, + "loss": 0.0085, + "step": 11890 + }, + { + "epoch": 17.894736842105264, + "grad_norm": 0.03832954540848732, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 11900 + }, + { + "epoch": 17.909774436090224, + "grad_norm": 0.06969650834798813, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11910 + }, + { + "epoch": 17.924812030075188, + "grad_norm": 0.049001529812812805, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 11920 + }, + { + "epoch": 17.93984962406015, + "grad_norm": 0.04849274456501007, + "learning_rate": 0.0005, + "loss": 0.008, + "step": 11930 + }, + { + "epoch": 17.954887218045112, + "grad_norm": 0.04725367948412895, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11940 + }, + { + "epoch": 17.969924812030076, + "grad_norm": 0.05629235878586769, + "learning_rate": 0.0005, + "loss": 0.0075, + "step": 11950 + }, + { + "epoch": 17.984962406015036, + "grad_norm": 0.0439433753490448, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11960 + }, + { + "epoch": 18.0, + "grad_norm": 0.04497876763343811, + "learning_rate": 0.0005, + "loss": 0.0078, + "step": 11970 + }, + { + "epoch": 18.015037593984964, + "grad_norm": 0.05537024140357971, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 11980 + }, + { + "epoch": 18.030075187969924, + "grad_norm": 0.04573548957705498, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 11990 + }, + { + "epoch": 18.045112781954888, + "grad_norm": 0.049514979124069214, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 12000 + }, + { + "epoch": 18.045112781954888, + "eval_cer": 0.02002682121051794, + "eval_loss": 0.07638879865407944, + "eval_runtime": 157.9735, + "eval_samples_per_second": 101.454, + "eval_steps_per_second": 0.798, + "eval_wer": 0.06803830179105039, + "step": 12000 + }, + { + "epoch": 18.06015037593985, + "grad_norm": 0.040032435208559036, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 12010 + }, + { + "epoch": 18.075187969924812, + "grad_norm": 0.04561951383948326, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 12020 + }, + { + "epoch": 18.090225563909776, + "grad_norm": 0.06437424570322037, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 12030 + }, + { + "epoch": 18.105263157894736, + "grad_norm": 0.08589992672204971, + "learning_rate": 0.0005, + "loss": 0.0062, + "step": 12040 + }, + { + "epoch": 18.1203007518797, + "grad_norm": 0.045134320855140686, + "learning_rate": 0.0005, + "loss": 0.0062, + "step": 12050 + }, + { + "epoch": 18.13533834586466, + "grad_norm": 0.04012113809585571, + "learning_rate": 0.0005, + "loss": 0.0063, + "step": 12060 + }, + { + "epoch": 18.150375939849624, + "grad_norm": 0.05758247524499893, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12070 + }, + { + "epoch": 18.165413533834588, + "grad_norm": 0.04539918154478073, + "learning_rate": 0.0005, + "loss": 0.0062, + "step": 12080 + }, + { + "epoch": 18.18045112781955, + "grad_norm": 0.040801793336868286, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 12090 + }, + { + "epoch": 18.195488721804512, + "grad_norm": 0.08883795142173767, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 12100 + }, + { + "epoch": 18.210526315789473, + "grad_norm": 0.043451208621263504, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 12110 + }, + { + "epoch": 18.225563909774436, + "grad_norm": 0.04162730649113655, + "learning_rate": 0.0005, + "loss": 0.0065, + "step": 12120 + }, + { + "epoch": 18.2406015037594, + "grad_norm": 0.04666702821850777, + "learning_rate": 0.0005, + "loss": 0.0065, + "step": 12130 + }, + { + "epoch": 18.25563909774436, + "grad_norm": 0.038976699113845825, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 12140 + }, + { + "epoch": 18.270676691729324, + "grad_norm": 0.09538313001394272, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 12150 + }, + { + "epoch": 18.285714285714285, + "grad_norm": 0.058226317167282104, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 12160 + }, + { + "epoch": 18.30075187969925, + "grad_norm": 0.06413419544696808, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 12170 + }, + { + "epoch": 18.31578947368421, + "grad_norm": 0.03986097872257233, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 12180 + }, + { + "epoch": 18.330827067669173, + "grad_norm": 0.044108662754297256, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 12190 + }, + { + "epoch": 18.345864661654137, + "grad_norm": 0.04044831916689873, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 12200 + }, + { + "epoch": 18.360902255639097, + "grad_norm": 0.044918570667505264, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 12210 + }, + { + "epoch": 18.37593984962406, + "grad_norm": 0.07513200491666794, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 12220 + }, + { + "epoch": 18.39097744360902, + "grad_norm": 0.06340505927801132, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 12230 + }, + { + "epoch": 18.406015037593985, + "grad_norm": 0.07325886189937592, + "learning_rate": 0.0005, + "loss": 0.0074, + "step": 12240 + }, + { + "epoch": 18.42105263157895, + "grad_norm": 0.051268450915813446, + "learning_rate": 0.0005, + "loss": 0.0066, + "step": 12250 + }, + { + "epoch": 18.43609022556391, + "grad_norm": 0.07912448048591614, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12260 + }, + { + "epoch": 18.451127819548873, + "grad_norm": 0.07773883640766144, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 12270 + }, + { + "epoch": 18.466165413533833, + "grad_norm": 0.05060628801584244, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12280 + }, + { + "epoch": 18.481203007518797, + "grad_norm": 0.040225569158792496, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 12290 + }, + { + "epoch": 18.49624060150376, + "grad_norm": 0.08059701323509216, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 12300 + }, + { + "epoch": 18.51127819548872, + "grad_norm": 0.061351656913757324, + "learning_rate": 0.0005, + "loss": 0.0068, + "step": 12310 + }, + { + "epoch": 18.526315789473685, + "grad_norm": 0.04331913962960243, + "learning_rate": 0.0005, + "loss": 0.0057, + "step": 12320 + }, + { + "epoch": 18.541353383458645, + "grad_norm": 0.07011712342500687, + "learning_rate": 0.0005, + "loss": 0.0066, + "step": 12330 + }, + { + "epoch": 18.55639097744361, + "grad_norm": 0.07536281645298004, + "learning_rate": 0.0005, + "loss": 0.007, + "step": 12340 + }, + { + "epoch": 18.571428571428573, + "grad_norm": 0.04379749298095703, + "learning_rate": 0.0005, + "loss": 0.0063, + "step": 12350 + }, + { + "epoch": 18.586466165413533, + "grad_norm": 0.06422626227140427, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 12360 + }, + { + "epoch": 18.601503759398497, + "grad_norm": 0.07083617895841599, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 12370 + }, + { + "epoch": 18.616541353383457, + "grad_norm": 0.03382309153676033, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 12380 + }, + { + "epoch": 18.63157894736842, + "grad_norm": 0.04558511823415756, + "learning_rate": 0.0005, + "loss": 0.0063, + "step": 12390 + }, + { + "epoch": 18.646616541353385, + "grad_norm": 0.037767112255096436, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12400 + }, + { + "epoch": 18.661654135338345, + "grad_norm": 0.0472201332449913, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 12410 + }, + { + "epoch": 18.67669172932331, + "grad_norm": 0.048948273062705994, + "learning_rate": 0.0005, + "loss": 0.0065, + "step": 12420 + }, + { + "epoch": 18.69172932330827, + "grad_norm": 0.07922033220529556, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 12430 + }, + { + "epoch": 18.706766917293233, + "grad_norm": 0.06186243146657944, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12440 + }, + { + "epoch": 18.721804511278194, + "grad_norm": 0.04407238960266113, + "learning_rate": 0.0005, + "loss": 0.0066, + "step": 12450 + }, + { + "epoch": 18.736842105263158, + "grad_norm": 0.051644254475831985, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 12460 + }, + { + "epoch": 18.75187969924812, + "grad_norm": 0.048314888030290604, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 12470 + }, + { + "epoch": 18.76691729323308, + "grad_norm": 0.04072916507720947, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 12480 + }, + { + "epoch": 18.781954887218046, + "grad_norm": 0.052623674273490906, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 12490 + }, + { + "epoch": 18.796992481203006, + "grad_norm": 0.05796734616160393, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 12500 + }, + { + "epoch": 18.81203007518797, + "grad_norm": 0.050005264580249786, + "learning_rate": 0.0005, + "loss": 0.0072, + "step": 12510 + }, + { + "epoch": 18.827067669172934, + "grad_norm": 0.04255378246307373, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12520 + }, + { + "epoch": 18.842105263157894, + "grad_norm": 0.06640315055847168, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 12530 + }, + { + "epoch": 18.857142857142858, + "grad_norm": 0.05466262623667717, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 12540 + }, + { + "epoch": 18.872180451127818, + "grad_norm": 0.053646743297576904, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 12550 + }, + { + "epoch": 18.887218045112782, + "grad_norm": 0.03561629354953766, + "learning_rate": 0.0005, + "loss": 0.0057, + "step": 12560 + }, + { + "epoch": 18.902255639097746, + "grad_norm": 0.04516750946640968, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12570 + }, + { + "epoch": 18.917293233082706, + "grad_norm": 0.0399198979139328, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 12580 + }, + { + "epoch": 18.93233082706767, + "grad_norm": 0.034706514328718185, + "learning_rate": 0.0005, + "loss": 0.0063, + "step": 12590 + }, + { + "epoch": 18.94736842105263, + "grad_norm": 0.08483397215604782, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 12600 + }, + { + "epoch": 18.962406015037594, + "grad_norm": 0.06673172861337662, + "learning_rate": 0.0005, + "loss": 0.0073, + "step": 12610 + }, + { + "epoch": 18.977443609022558, + "grad_norm": 0.0412411242723465, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 12620 + }, + { + "epoch": 18.992481203007518, + "grad_norm": 0.03789719194173813, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 12630 + }, + { + "epoch": 19.007518796992482, + "grad_norm": 0.043958283960819244, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 12640 + }, + { + "epoch": 19.022556390977442, + "grad_norm": 0.03598282113671303, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 12650 + }, + { + "epoch": 19.037593984962406, + "grad_norm": 0.05897130072116852, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 12660 + }, + { + "epoch": 19.05263157894737, + "grad_norm": 0.033550869673490524, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 12670 + }, + { + "epoch": 19.06766917293233, + "grad_norm": 0.10478643327951431, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12680 + }, + { + "epoch": 19.082706766917294, + "grad_norm": 0.039542168378829956, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 12690 + }, + { + "epoch": 19.097744360902254, + "grad_norm": 0.0462321899831295, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 12700 + }, + { + "epoch": 19.112781954887218, + "grad_norm": 0.04623578488826752, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 12710 + }, + { + "epoch": 19.127819548872182, + "grad_norm": 0.055180490016937256, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 12720 + }, + { + "epoch": 19.142857142857142, + "grad_norm": 0.03973427414894104, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 12730 + }, + { + "epoch": 19.157894736842106, + "grad_norm": 0.1164124459028244, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 12740 + }, + { + "epoch": 19.172932330827066, + "grad_norm": 0.0388130359351635, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 12750 + }, + { + "epoch": 19.18796992481203, + "grad_norm": 0.04107864946126938, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 12760 + }, + { + "epoch": 19.203007518796994, + "grad_norm": 0.04334622621536255, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12770 + }, + { + "epoch": 19.218045112781954, + "grad_norm": 0.0662887841463089, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 12780 + }, + { + "epoch": 19.23308270676692, + "grad_norm": 0.04289138689637184, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 12790 + }, + { + "epoch": 19.24812030075188, + "grad_norm": 0.04197674244642258, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 12800 + }, + { + "epoch": 19.263157894736842, + "grad_norm": 0.03499612212181091, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12810 + }, + { + "epoch": 19.278195488721803, + "grad_norm": 0.04472319409251213, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12820 + }, + { + "epoch": 19.293233082706767, + "grad_norm": 0.041913941502571106, + "learning_rate": 0.0005, + "loss": 0.0057, + "step": 12830 + }, + { + "epoch": 19.30827067669173, + "grad_norm": 0.05282393842935562, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 12840 + }, + { + "epoch": 19.32330827067669, + "grad_norm": 0.04793861135840416, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 12850 + }, + { + "epoch": 19.338345864661655, + "grad_norm": 0.04471307992935181, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 12860 + }, + { + "epoch": 19.353383458646615, + "grad_norm": 0.08557348698377609, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 12870 + }, + { + "epoch": 19.36842105263158, + "grad_norm": 0.039792515337467194, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 12880 + }, + { + "epoch": 19.383458646616543, + "grad_norm": 0.03936640918254852, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12890 + }, + { + "epoch": 19.398496240601503, + "grad_norm": 0.04597417265176773, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12900 + }, + { + "epoch": 19.413533834586467, + "grad_norm": 0.04913599416613579, + "learning_rate": 0.0005, + "loss": 0.0062, + "step": 12910 + }, + { + "epoch": 19.428571428571427, + "grad_norm": 0.03916778042912483, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 12920 + }, + { + "epoch": 19.44360902255639, + "grad_norm": 0.041145097464323044, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 12930 + }, + { + "epoch": 19.458646616541355, + "grad_norm": 0.044809699058532715, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 12940 + }, + { + "epoch": 19.473684210526315, + "grad_norm": 0.06783033907413483, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12950 + }, + { + "epoch": 19.48872180451128, + "grad_norm": 0.050151437520980835, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 12960 + }, + { + "epoch": 19.50375939849624, + "grad_norm": 0.040898412466049194, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12970 + }, + { + "epoch": 19.518796992481203, + "grad_norm": 0.056583017110824585, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 12980 + }, + { + "epoch": 19.533834586466167, + "grad_norm": 0.05618445202708244, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 12990 + }, + { + "epoch": 19.548872180451127, + "grad_norm": 0.04943498969078064, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 13000 + }, + { + "epoch": 19.548872180451127, + "eval_cer": 0.020165551609748665, + "eval_loss": 0.07748271524906158, + "eval_runtime": 157.7447, + "eval_samples_per_second": 101.601, + "eval_steps_per_second": 0.799, + "eval_wer": 0.06873367287488959, + "step": 13000 + }, + { + "epoch": 19.56390977443609, + "grad_norm": 0.04475477710366249, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 13010 + }, + { + "epoch": 19.57894736842105, + "grad_norm": 0.05696471035480499, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13020 + }, + { + "epoch": 19.593984962406015, + "grad_norm": 0.048657506704330444, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 13030 + }, + { + "epoch": 19.60902255639098, + "grad_norm": 0.04403107240796089, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 13040 + }, + { + "epoch": 19.62406015037594, + "grad_norm": 0.06767864525318146, + "learning_rate": 0.0005, + "loss": 0.0057, + "step": 13050 + }, + { + "epoch": 19.639097744360903, + "grad_norm": 0.06452973932027817, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 13060 + }, + { + "epoch": 19.654135338345863, + "grad_norm": 0.047971490770578384, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 13070 + }, + { + "epoch": 19.669172932330827, + "grad_norm": 0.031981468200683594, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13080 + }, + { + "epoch": 19.68421052631579, + "grad_norm": 0.08796112984418869, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 13090 + }, + { + "epoch": 19.69924812030075, + "grad_norm": 0.06189596280455589, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13100 + }, + { + "epoch": 19.714285714285715, + "grad_norm": 0.047975972294807434, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 13110 + }, + { + "epoch": 19.729323308270676, + "grad_norm": 0.04918560013175011, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 13120 + }, + { + "epoch": 19.74436090225564, + "grad_norm": 0.04995375871658325, + "learning_rate": 0.0005, + "loss": 0.0057, + "step": 13130 + }, + { + "epoch": 19.7593984962406, + "grad_norm": 0.047466423362493515, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 13140 + }, + { + "epoch": 19.774436090225564, + "grad_norm": 0.055160801857709885, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 13150 + }, + { + "epoch": 19.789473684210527, + "grad_norm": 0.03777928650379181, + "learning_rate": 0.0005, + "loss": 0.0069, + "step": 13160 + }, + { + "epoch": 19.804511278195488, + "grad_norm": 0.06801455467939377, + "learning_rate": 0.0005, + "loss": 0.0066, + "step": 13170 + }, + { + "epoch": 19.81954887218045, + "grad_norm": 0.03702409565448761, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 13180 + }, + { + "epoch": 19.834586466165412, + "grad_norm": 0.07816283404827118, + "learning_rate": 0.0005, + "loss": 0.0067, + "step": 13190 + }, + { + "epoch": 19.849624060150376, + "grad_norm": 0.057374462485313416, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 13200 + }, + { + "epoch": 19.86466165413534, + "grad_norm": 0.03745908662676811, + "learning_rate": 0.0005, + "loss": 0.0064, + "step": 13210 + }, + { + "epoch": 19.8796992481203, + "grad_norm": 0.03968452289700508, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 13220 + }, + { + "epoch": 19.894736842105264, + "grad_norm": 0.058864228427410126, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 13230 + }, + { + "epoch": 19.909774436090224, + "grad_norm": 0.055964235216379166, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 13240 + }, + { + "epoch": 19.924812030075188, + "grad_norm": 0.0497111938893795, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 13250 + }, + { + "epoch": 19.93984962406015, + "grad_norm": 0.049517713487148285, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 13260 + }, + { + "epoch": 19.954887218045112, + "grad_norm": 0.04816407337784767, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13270 + }, + { + "epoch": 19.969924812030076, + "grad_norm": 0.04329106584191322, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 13280 + }, + { + "epoch": 19.984962406015036, + "grad_norm": 0.04538606107234955, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 13290 + }, + { + "epoch": 20.0, + "grad_norm": 0.05977080389857292, + "learning_rate": 0.0005, + "loss": 0.0066, + "step": 13300 + }, + { + "epoch": 20.015037593984964, + "grad_norm": 0.04326824098825455, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 13310 + }, + { + "epoch": 20.030075187969924, + "grad_norm": 0.07028619199991226, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13320 + }, + { + "epoch": 20.045112781954888, + "grad_norm": 0.043210115283727646, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 13330 + }, + { + "epoch": 20.06015037593985, + "grad_norm": 0.0579630583524704, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13340 + }, + { + "epoch": 20.075187969924812, + "grad_norm": 0.04393903911113739, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 13350 + }, + { + "epoch": 20.090225563909776, + "grad_norm": 0.036538902670145035, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13360 + }, + { + "epoch": 20.105263157894736, + "grad_norm": 0.04232252761721611, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 13370 + }, + { + "epoch": 20.1203007518797, + "grad_norm": 0.04825455695390701, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 13380 + }, + { + "epoch": 20.13533834586466, + "grad_norm": 0.045079343020915985, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 13390 + }, + { + "epoch": 20.150375939849624, + "grad_norm": 0.07188252359628677, + "learning_rate": 0.0005, + "loss": 0.0061, + "step": 13400 + }, + { + "epoch": 20.165413533834588, + "grad_norm": 0.04869913309812546, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13410 + }, + { + "epoch": 20.18045112781955, + "grad_norm": 0.03089517541229725, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 13420 + }, + { + "epoch": 20.195488721804512, + "grad_norm": 0.04251734912395477, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 13430 + }, + { + "epoch": 20.210526315789473, + "grad_norm": 0.0468149408698082, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13440 + }, + { + "epoch": 20.225563909774436, + "grad_norm": 0.055634401738643646, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 13450 + }, + { + "epoch": 20.2406015037594, + "grad_norm": 0.048135533928871155, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13460 + }, + { + "epoch": 20.25563909774436, + "grad_norm": 0.054604172706604004, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 13470 + }, + { + "epoch": 20.270676691729324, + "grad_norm": 0.06354694068431854, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 13480 + }, + { + "epoch": 20.285714285714285, + "grad_norm": 0.03915797919034958, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 13490 + }, + { + "epoch": 20.30075187969925, + "grad_norm": 0.03620188310742378, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13500 + }, + { + "epoch": 20.31578947368421, + "grad_norm": 0.05438293144106865, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 13510 + }, + { + "epoch": 20.330827067669173, + "grad_norm": 0.03266293555498123, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13520 + }, + { + "epoch": 20.345864661654137, + "grad_norm": 0.04873530566692352, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 13530 + }, + { + "epoch": 20.360902255639097, + "grad_norm": 0.03773298114538193, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 13540 + }, + { + "epoch": 20.37593984962406, + "grad_norm": 0.054863449186086655, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 13550 + }, + { + "epoch": 20.39097744360902, + "grad_norm": 0.041172876954078674, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13560 + }, + { + "epoch": 20.406015037593985, + "grad_norm": 0.053773295134305954, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 13570 + }, + { + "epoch": 20.42105263157895, + "grad_norm": 0.05141289904713631, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 13580 + }, + { + "epoch": 20.43609022556391, + "grad_norm": 0.05964859202504158, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13590 + }, + { + "epoch": 20.451127819548873, + "grad_norm": 0.04413893073797226, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 13600 + }, + { + "epoch": 20.466165413533833, + "grad_norm": 0.04389195889234543, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13610 + }, + { + "epoch": 20.481203007518797, + "grad_norm": 0.06117071956396103, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13620 + }, + { + "epoch": 20.49624060150376, + "grad_norm": 0.04524886980652809, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13630 + }, + { + "epoch": 20.51127819548872, + "grad_norm": 0.066983662545681, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13640 + }, + { + "epoch": 20.526315789473685, + "grad_norm": 0.03515216335654259, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 13650 + }, + { + "epoch": 20.541353383458645, + "grad_norm": 0.05180105194449425, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 13660 + }, + { + "epoch": 20.55639097744361, + "grad_norm": 0.03179685398936272, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13670 + }, + { + "epoch": 20.571428571428573, + "grad_norm": 0.036657001823186874, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 13680 + }, + { + "epoch": 20.586466165413533, + "grad_norm": 0.03163471072912216, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 13690 + }, + { + "epoch": 20.601503759398497, + "grad_norm": 0.07313386350870132, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 13700 + }, + { + "epoch": 20.616541353383457, + "grad_norm": 0.03774179518222809, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 13710 + }, + { + "epoch": 20.63157894736842, + "grad_norm": 0.055352889001369476, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 13720 + }, + { + "epoch": 20.646616541353385, + "grad_norm": 0.05922696739435196, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 13730 + }, + { + "epoch": 20.661654135338345, + "grad_norm": 0.020613932982087135, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13740 + }, + { + "epoch": 20.67669172932331, + "grad_norm": 0.02521524392068386, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13750 + }, + { + "epoch": 20.69172932330827, + "grad_norm": 0.053172748535871506, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13760 + }, + { + "epoch": 20.706766917293233, + "grad_norm": 0.060833852738142014, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13770 + }, + { + "epoch": 20.721804511278194, + "grad_norm": 0.043886635452508926, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 13780 + }, + { + "epoch": 20.736842105263158, + "grad_norm": 0.061435725539922714, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13790 + }, + { + "epoch": 20.75187969924812, + "grad_norm": 0.05407465621829033, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 13800 + }, + { + "epoch": 20.76691729323308, + "grad_norm": 0.07196366041898727, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13810 + }, + { + "epoch": 20.781954887218046, + "grad_norm": 0.050324875861406326, + "learning_rate": 0.0005, + "loss": 0.0057, + "step": 13820 + }, + { + "epoch": 20.796992481203006, + "grad_norm": 0.06420844048261642, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13830 + }, + { + "epoch": 20.81203007518797, + "grad_norm": 0.03528842329978943, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13840 + }, + { + "epoch": 20.827067669172934, + "grad_norm": 0.044911082834005356, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13850 + }, + { + "epoch": 20.842105263157894, + "grad_norm": 0.03576183691620827, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 13860 + }, + { + "epoch": 20.857142857142858, + "grad_norm": 0.04545877128839493, + "learning_rate": 0.0005, + "loss": 0.0058, + "step": 13870 + }, + { + "epoch": 20.872180451127818, + "grad_norm": 0.03951716050505638, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 13880 + }, + { + "epoch": 20.887218045112782, + "grad_norm": 0.049765873700380325, + "learning_rate": 0.0005, + "loss": 0.0059, + "step": 13890 + }, + { + "epoch": 20.902255639097746, + "grad_norm": 0.04492948204278946, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 13900 + }, + { + "epoch": 20.917293233082706, + "grad_norm": 0.06116321682929993, + "learning_rate": 0.0005, + "loss": 0.006, + "step": 13910 + }, + { + "epoch": 20.93233082706767, + "grad_norm": 0.059679917991161346, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 13920 + }, + { + "epoch": 20.94736842105263, + "grad_norm": 0.03961421921849251, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 13930 + }, + { + "epoch": 20.962406015037594, + "grad_norm": 0.03671489283442497, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 13940 + }, + { + "epoch": 20.977443609022558, + "grad_norm": 0.03708609193563461, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 13950 + }, + { + "epoch": 20.992481203007518, + "grad_norm": 0.033162325620651245, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 13960 + }, + { + "epoch": 21.007518796992482, + "grad_norm": 0.026664569973945618, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 13970 + }, + { + "epoch": 21.022556390977442, + "grad_norm": 0.045722655951976776, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 13980 + }, + { + "epoch": 21.037593984962406, + "grad_norm": 0.03403405845165253, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 13990 + }, + { + "epoch": 21.05263157894737, + "grad_norm": 0.07509675621986389, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14000 + }, + { + "epoch": 21.05263157894737, + "eval_cer": 0.019106032188172827, + "eval_loss": 0.08213265985250473, + "eval_runtime": 158.2029, + "eval_samples_per_second": 101.307, + "eval_steps_per_second": 0.796, + "eval_wer": 0.06935856715969103, + "step": 14000 + }, + { + "epoch": 21.06766917293233, + "grad_norm": 0.048846278339624405, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 14010 + }, + { + "epoch": 21.082706766917294, + "grad_norm": 0.046284522861242294, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14020 + }, + { + "epoch": 21.097744360902254, + "grad_norm": 0.03692031279206276, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14030 + }, + { + "epoch": 21.112781954887218, + "grad_norm": 0.044887643307447433, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 14040 + }, + { + "epoch": 21.127819548872182, + "grad_norm": 0.04559887945652008, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 14050 + }, + { + "epoch": 21.142857142857142, + "grad_norm": 0.04804844409227371, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14060 + }, + { + "epoch": 21.157894736842106, + "grad_norm": 0.04476190358400345, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 14070 + }, + { + "epoch": 21.172932330827066, + "grad_norm": 0.04622158408164978, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 14080 + }, + { + "epoch": 21.18796992481203, + "grad_norm": 0.06559256464242935, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 14090 + }, + { + "epoch": 21.203007518796994, + "grad_norm": 0.05269289389252663, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14100 + }, + { + "epoch": 21.218045112781954, + "grad_norm": 0.038871847093105316, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 14110 + }, + { + "epoch": 21.23308270676692, + "grad_norm": 0.04512747749686241, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 14120 + }, + { + "epoch": 21.24812030075188, + "grad_norm": 0.034506604075431824, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 14130 + }, + { + "epoch": 21.263157894736842, + "grad_norm": 0.06643722206354141, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 14140 + }, + { + "epoch": 21.278195488721803, + "grad_norm": 0.05581054836511612, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14150 + }, + { + "epoch": 21.293233082706767, + "grad_norm": 0.050910234451293945, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 14160 + }, + { + "epoch": 21.30827067669173, + "grad_norm": 0.04692579060792923, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14170 + }, + { + "epoch": 21.32330827067669, + "grad_norm": 0.05351801589131355, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14180 + }, + { + "epoch": 21.338345864661655, + "grad_norm": 0.03689511492848396, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 14190 + }, + { + "epoch": 21.353383458646615, + "grad_norm": 0.03011569194495678, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 14200 + }, + { + "epoch": 21.36842105263158, + "grad_norm": 0.04333231970667839, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 14210 + }, + { + "epoch": 21.383458646616543, + "grad_norm": 0.040374305099248886, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 14220 + }, + { + "epoch": 21.398496240601503, + "grad_norm": 0.037551503628492355, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 14230 + }, + { + "epoch": 21.413533834586467, + "grad_norm": 0.05253462493419647, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14240 + }, + { + "epoch": 21.428571428571427, + "grad_norm": 0.05089132860302925, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14250 + }, + { + "epoch": 21.44360902255639, + "grad_norm": 0.03457905352115631, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 14260 + }, + { + "epoch": 21.458646616541355, + "grad_norm": 0.058625366538763046, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 14270 + }, + { + "epoch": 21.473684210526315, + "grad_norm": 0.043307531625032425, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 14280 + }, + { + "epoch": 21.48872180451128, + "grad_norm": 0.04146287590265274, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14290 + }, + { + "epoch": 21.50375939849624, + "grad_norm": 0.056128159165382385, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14300 + }, + { + "epoch": 21.518796992481203, + "grad_norm": 0.06470511853694916, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14310 + }, + { + "epoch": 21.533834586466167, + "grad_norm": 0.039793167263269424, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 14320 + }, + { + "epoch": 21.548872180451127, + "grad_norm": 0.035141363739967346, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 14330 + }, + { + "epoch": 21.56390977443609, + "grad_norm": 0.054246947169303894, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 14340 + }, + { + "epoch": 21.57894736842105, + "grad_norm": 0.04859871044754982, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 14350 + }, + { + "epoch": 21.593984962406015, + "grad_norm": 0.03320176154375076, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14360 + }, + { + "epoch": 21.60902255639098, + "grad_norm": 0.04123358428478241, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14370 + }, + { + "epoch": 21.62406015037594, + "grad_norm": 0.06261591613292694, + "learning_rate": 0.0005, + "loss": 0.0051, + "step": 14380 + }, + { + "epoch": 21.639097744360903, + "grad_norm": 0.03668075427412987, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 14390 + }, + { + "epoch": 21.654135338345863, + "grad_norm": 0.03023553267121315, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 14400 + }, + { + "epoch": 21.669172932330827, + "grad_norm": 0.04237750545144081, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 14410 + }, + { + "epoch": 21.68421052631579, + "grad_norm": 0.03962303698062897, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14420 + }, + { + "epoch": 21.69924812030075, + "grad_norm": 0.05062340199947357, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14430 + }, + { + "epoch": 21.714285714285715, + "grad_norm": 0.03545669838786125, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 14440 + }, + { + "epoch": 21.729323308270676, + "grad_norm": 0.05113904923200607, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 14450 + }, + { + "epoch": 21.74436090225564, + "grad_norm": 0.039374858140945435, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14460 + }, + { + "epoch": 21.7593984962406, + "grad_norm": 0.05903002619743347, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14470 + }, + { + "epoch": 21.774436090225564, + "grad_norm": 0.03725092485547066, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14480 + }, + { + "epoch": 21.789473684210527, + "grad_norm": 0.05206378921866417, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14490 + }, + { + "epoch": 21.804511278195488, + "grad_norm": 0.05254136025905609, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14500 + }, + { + "epoch": 21.81954887218045, + "grad_norm": 0.035321369767189026, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14510 + }, + { + "epoch": 21.834586466165412, + "grad_norm": 0.047776296734809875, + "learning_rate": 0.0005, + "loss": 0.0056, + "step": 14520 + }, + { + "epoch": 21.849624060150376, + "grad_norm": 0.06352993100881577, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14530 + }, + { + "epoch": 21.86466165413534, + "grad_norm": 0.04481920599937439, + "learning_rate": 0.0005, + "loss": 0.0055, + "step": 14540 + }, + { + "epoch": 21.8796992481203, + "grad_norm": 0.04325064644217491, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14550 + }, + { + "epoch": 21.894736842105264, + "grad_norm": 0.04412194341421127, + "learning_rate": 0.0005, + "loss": 0.0054, + "step": 14560 + }, + { + "epoch": 21.909774436090224, + "grad_norm": 0.04136514663696289, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 14570 + }, + { + "epoch": 21.924812030075188, + "grad_norm": 0.060158200562000275, + "learning_rate": 0.0005, + "loss": 0.005, + "step": 14580 + }, + { + "epoch": 21.93984962406015, + "grad_norm": 0.048980843275785446, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14590 + }, + { + "epoch": 21.954887218045112, + "grad_norm": 0.03913483768701553, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 14600 + }, + { + "epoch": 21.969924812030076, + "grad_norm": 0.03526630625128746, + "learning_rate": 0.0005, + "loss": 0.0053, + "step": 14610 + }, + { + "epoch": 21.984962406015036, + "grad_norm": 0.04283861815929413, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14620 + }, + { + "epoch": 22.0, + "grad_norm": 0.05301738902926445, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 14630 + }, + { + "epoch": 22.015037593984964, + "grad_norm": 0.037607382982969284, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 14640 + }, + { + "epoch": 22.030075187969924, + "grad_norm": 0.03292655572295189, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 14650 + }, + { + "epoch": 22.045112781954888, + "grad_norm": 0.05127028375864029, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 14660 + }, + { + "epoch": 22.06015037593985, + "grad_norm": 0.02501407265663147, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 14670 + }, + { + "epoch": 22.075187969924812, + "grad_norm": 0.03535280376672745, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 14680 + }, + { + "epoch": 22.090225563909776, + "grad_norm": 0.03969969600439072, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 14690 + }, + { + "epoch": 22.105263157894736, + "grad_norm": 0.05119691044092178, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 14700 + }, + { + "epoch": 22.1203007518797, + "grad_norm": 0.03322632983326912, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 14710 + }, + { + "epoch": 22.13533834586466, + "grad_norm": 0.03751862794160843, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 14720 + }, + { + "epoch": 22.150375939849624, + "grad_norm": 0.03397894278168678, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 14730 + }, + { + "epoch": 22.165413533834588, + "grad_norm": 0.06201380118727684, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14740 + }, + { + "epoch": 22.18045112781955, + "grad_norm": 0.034275736659765244, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14750 + }, + { + "epoch": 22.195488721804512, + "grad_norm": 0.04140623286366463, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14760 + }, + { + "epoch": 22.210526315789473, + "grad_norm": 0.029766838997602463, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 14770 + }, + { + "epoch": 22.225563909774436, + "grad_norm": 0.04038533940911293, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14780 + }, + { + "epoch": 22.2406015037594, + "grad_norm": 0.04586336016654968, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 14790 + }, + { + "epoch": 22.25563909774436, + "grad_norm": 0.0663742795586586, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14800 + }, + { + "epoch": 22.270676691729324, + "grad_norm": 0.07037439942359924, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 14810 + }, + { + "epoch": 22.285714285714285, + "grad_norm": 0.04160144925117493, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14820 + }, + { + "epoch": 22.30075187969925, + "grad_norm": 0.04149804636836052, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 14830 + }, + { + "epoch": 22.31578947368421, + "grad_norm": 0.040650028735399246, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14840 + }, + { + "epoch": 22.330827067669173, + "grad_norm": 0.04660551995038986, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 14850 + }, + { + "epoch": 22.345864661654137, + "grad_norm": 0.05716900900006294, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14860 + }, + { + "epoch": 22.360902255639097, + "grad_norm": 0.040852632373571396, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14870 + }, + { + "epoch": 22.37593984962406, + "grad_norm": 0.03741825371980667, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14880 + }, + { + "epoch": 22.39097744360902, + "grad_norm": 0.03582283854484558, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 14890 + }, + { + "epoch": 22.406015037593985, + "grad_norm": 0.054151877760887146, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14900 + }, + { + "epoch": 22.42105263157895, + "grad_norm": 0.03782195225358009, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 14910 + }, + { + "epoch": 22.43609022556391, + "grad_norm": 0.05374496430158615, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 14920 + }, + { + "epoch": 22.451127819548873, + "grad_norm": 0.05969220772385597, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 14930 + }, + { + "epoch": 22.466165413533833, + "grad_norm": 0.03430042788386345, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 14940 + }, + { + "epoch": 22.481203007518797, + "grad_norm": 0.029958875849843025, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 14950 + }, + { + "epoch": 22.49624060150376, + "grad_norm": 0.038399506360292435, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 14960 + }, + { + "epoch": 22.51127819548872, + "grad_norm": 0.03843643143773079, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 14970 + }, + { + "epoch": 22.526315789473685, + "grad_norm": 0.06815985590219498, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 14980 + }, + { + "epoch": 22.541353383458645, + "grad_norm": 0.04166082292795181, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 14990 + }, + { + "epoch": 22.55639097744361, + "grad_norm": 0.040058065205812454, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 15000 + }, + { + "epoch": 22.55639097744361, + "eval_cer": 0.01970515709857611, + "eval_loss": 0.08506421744823456, + "eval_runtime": 159.1662, + "eval_samples_per_second": 100.693, + "eval_steps_per_second": 0.792, + "eval_wer": 0.06924110582796143, + "step": 15000 + }, + { + "epoch": 22.571428571428573, + "grad_norm": 0.056909333914518356, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 15010 + }, + { + "epoch": 22.586466165413533, + "grad_norm": 0.04462490603327751, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15020 + }, + { + "epoch": 22.601503759398497, + "grad_norm": 0.035606950521469116, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15030 + }, + { + "epoch": 22.616541353383457, + "grad_norm": 0.0491686575114727, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15040 + }, + { + "epoch": 22.63157894736842, + "grad_norm": 0.0657297670841217, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 15050 + }, + { + "epoch": 22.646616541353385, + "grad_norm": 0.030945518985390663, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15060 + }, + { + "epoch": 22.661654135338345, + "grad_norm": 0.05141136795282364, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 15070 + }, + { + "epoch": 22.67669172932331, + "grad_norm": 0.06340225040912628, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 15080 + }, + { + "epoch": 22.69172932330827, + "grad_norm": 0.04180299490690231, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15090 + }, + { + "epoch": 22.706766917293233, + "grad_norm": 0.039200399070978165, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15100 + }, + { + "epoch": 22.721804511278194, + "grad_norm": 0.04693165421485901, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 15110 + }, + { + "epoch": 22.736842105263158, + "grad_norm": 0.05959663912653923, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15120 + }, + { + "epoch": 22.75187969924812, + "grad_norm": 0.0449870266020298, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15130 + }, + { + "epoch": 22.76691729323308, + "grad_norm": 0.05562209710478783, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 15140 + }, + { + "epoch": 22.781954887218046, + "grad_norm": 0.041493240743875504, + "learning_rate": 0.0005, + "loss": 0.0052, + "step": 15150 + }, + { + "epoch": 22.796992481203006, + "grad_norm": 0.03672722727060318, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15160 + }, + { + "epoch": 22.81203007518797, + "grad_norm": 0.03829977661371231, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15170 + }, + { + "epoch": 22.827067669172934, + "grad_norm": 0.041138771921396255, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 15180 + }, + { + "epoch": 22.842105263157894, + "grad_norm": 0.04923051595687866, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15190 + }, + { + "epoch": 22.857142857142858, + "grad_norm": 0.05512337014079094, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15200 + }, + { + "epoch": 22.872180451127818, + "grad_norm": 0.03917677327990532, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 15210 + }, + { + "epoch": 22.887218045112782, + "grad_norm": 0.04397207498550415, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 15220 + }, + { + "epoch": 22.902255639097746, + "grad_norm": 0.030948463827371597, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15230 + }, + { + "epoch": 22.917293233082706, + "grad_norm": 0.03791814297437668, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15240 + }, + { + "epoch": 22.93233082706767, + "grad_norm": 0.049366023391485214, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 15250 + }, + { + "epoch": 22.94736842105263, + "grad_norm": 0.046063438057899475, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 15260 + }, + { + "epoch": 22.962406015037594, + "grad_norm": 0.04583323374390602, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 15270 + }, + { + "epoch": 22.977443609022558, + "grad_norm": 0.03475106880068779, + "learning_rate": 0.0005, + "loss": 0.0049, + "step": 15280 + }, + { + "epoch": 22.992481203007518, + "grad_norm": 0.032114483416080475, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15290 + }, + { + "epoch": 23.007518796992482, + "grad_norm": 0.04077647253870964, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15300 + }, + { + "epoch": 23.022556390977442, + "grad_norm": 0.026594066992402077, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 15310 + }, + { + "epoch": 23.037593984962406, + "grad_norm": 0.050328269600868225, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15320 + }, + { + "epoch": 23.05263157894737, + "grad_norm": 0.03770780563354492, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 15330 + }, + { + "epoch": 23.06766917293233, + "grad_norm": 0.08038242161273956, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 15340 + }, + { + "epoch": 23.082706766917294, + "grad_norm": 0.05860363692045212, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15350 + }, + { + "epoch": 23.097744360902254, + "grad_norm": 0.050166256725788116, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15360 + }, + { + "epoch": 23.112781954887218, + "grad_norm": 0.03364124149084091, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15370 + }, + { + "epoch": 23.127819548872182, + "grad_norm": 0.03216736763715744, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15380 + }, + { + "epoch": 23.142857142857142, + "grad_norm": 0.02895667403936386, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 15390 + }, + { + "epoch": 23.157894736842106, + "grad_norm": 0.06881210952997208, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15400 + }, + { + "epoch": 23.172932330827066, + "grad_norm": 0.05353430286049843, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 15410 + }, + { + "epoch": 23.18796992481203, + "grad_norm": 0.05390486493706703, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15420 + }, + { + "epoch": 23.203007518796994, + "grad_norm": 0.03650097921490669, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15430 + }, + { + "epoch": 23.218045112781954, + "grad_norm": 0.0385526642203331, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15440 + }, + { + "epoch": 23.23308270676692, + "grad_norm": 0.04764069244265556, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 15450 + }, + { + "epoch": 23.24812030075188, + "grad_norm": 0.03533631190657616, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 15460 + }, + { + "epoch": 23.263157894736842, + "grad_norm": 0.04487557336688042, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 15470 + }, + { + "epoch": 23.278195488721803, + "grad_norm": 0.03154211491346359, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 15480 + }, + { + "epoch": 23.293233082706767, + "grad_norm": 0.047045398503541946, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15490 + }, + { + "epoch": 23.30827067669173, + "grad_norm": 0.028156524524092674, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 15500 + }, + { + "epoch": 23.32330827067669, + "grad_norm": 0.0352107509970665, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15510 + }, + { + "epoch": 23.338345864661655, + "grad_norm": 0.038604650646448135, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15520 + }, + { + "epoch": 23.353383458646615, + "grad_norm": 0.05063849315047264, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 15530 + }, + { + "epoch": 23.36842105263158, + "grad_norm": 0.043553225696086884, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 15540 + }, + { + "epoch": 23.383458646616543, + "grad_norm": 0.02239191345870495, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15550 + }, + { + "epoch": 23.398496240601503, + "grad_norm": 0.04093382507562637, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 15560 + }, + { + "epoch": 23.413533834586467, + "grad_norm": 0.06343573331832886, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15570 + }, + { + "epoch": 23.428571428571427, + "grad_norm": 0.05281996354460716, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 15580 + }, + { + "epoch": 23.44360902255639, + "grad_norm": 0.03988668695092201, + "learning_rate": 0.0005, + "loss": 0.0047, + "step": 15590 + }, + { + "epoch": 23.458646616541355, + "grad_norm": 0.04888569191098213, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15600 + }, + { + "epoch": 23.473684210526315, + "grad_norm": 0.04654020816087723, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 15610 + }, + { + "epoch": 23.48872180451128, + "grad_norm": 0.040350932627916336, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15620 + }, + { + "epoch": 23.50375939849624, + "grad_norm": 0.03500589728355408, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15630 + }, + { + "epoch": 23.518796992481203, + "grad_norm": 0.04560462012887001, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 15640 + }, + { + "epoch": 23.533834586466167, + "grad_norm": 0.05435361713171005, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15650 + }, + { + "epoch": 23.548872180451127, + "grad_norm": 0.04567120224237442, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 15660 + }, + { + "epoch": 23.56390977443609, + "grad_norm": 0.036758434027433395, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 15670 + }, + { + "epoch": 23.57894736842105, + "grad_norm": 0.04143821448087692, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15680 + }, + { + "epoch": 23.593984962406015, + "grad_norm": 0.033227793872356415, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15690 + }, + { + "epoch": 23.60902255639098, + "grad_norm": 0.03904704377055168, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15700 + }, + { + "epoch": 23.62406015037594, + "grad_norm": 0.05013486370444298, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15710 + }, + { + "epoch": 23.639097744360903, + "grad_norm": 0.04018151015043259, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 15720 + }, + { + "epoch": 23.654135338345863, + "grad_norm": 0.0354878231883049, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15730 + }, + { + "epoch": 23.669172932330827, + "grad_norm": 0.04210834577679634, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15740 + }, + { + "epoch": 23.68421052631579, + "grad_norm": 0.03412095829844475, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 15750 + }, + { + "epoch": 23.69924812030075, + "grad_norm": 0.03264550492167473, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15760 + }, + { + "epoch": 23.714285714285715, + "grad_norm": 0.03353846073150635, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15770 + }, + { + "epoch": 23.729323308270676, + "grad_norm": 0.03808584064245224, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 15780 + }, + { + "epoch": 23.74436090225564, + "grad_norm": 0.03541106358170509, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 15790 + }, + { + "epoch": 23.7593984962406, + "grad_norm": 0.10308961570262909, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 15800 + }, + { + "epoch": 23.774436090225564, + "grad_norm": 0.03687261790037155, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 15810 + }, + { + "epoch": 23.789473684210527, + "grad_norm": 0.04498017579317093, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15820 + }, + { + "epoch": 23.804511278195488, + "grad_norm": 0.04334450885653496, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15830 + }, + { + "epoch": 23.81954887218045, + "grad_norm": 0.04203008860349655, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 15840 + }, + { + "epoch": 23.834586466165412, + "grad_norm": 0.03529461473226547, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15850 + }, + { + "epoch": 23.849624060150376, + "grad_norm": 0.05043353512883186, + "learning_rate": 0.0005, + "loss": 0.0046, + "step": 15860 + }, + { + "epoch": 23.86466165413534, + "grad_norm": 0.03910822793841362, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 15870 + }, + { + "epoch": 23.8796992481203, + "grad_norm": 0.04063506796956062, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 15880 + }, + { + "epoch": 23.894736842105264, + "grad_norm": 0.05736089497804642, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 15890 + }, + { + "epoch": 23.909774436090224, + "grad_norm": 0.053351107984781265, + "learning_rate": 0.0005, + "loss": 0.0048, + "step": 15900 + }, + { + "epoch": 23.924812030075188, + "grad_norm": 0.03433650732040405, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 15910 + }, + { + "epoch": 23.93984962406015, + "grad_norm": 0.040141090750694275, + "learning_rate": 0.0005, + "loss": 0.0044, + "step": 15920 + }, + { + "epoch": 23.954887218045112, + "grad_norm": 0.04697522148489952, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 15930 + }, + { + "epoch": 23.969924812030076, + "grad_norm": 0.0393383614718914, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 15940 + }, + { + "epoch": 23.984962406015036, + "grad_norm": 0.04050549119710922, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 15950 + }, + { + "epoch": 24.0, + "grad_norm": 0.05381736159324646, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 15960 + }, + { + "epoch": 24.015037593984964, + "grad_norm": 0.03507491573691368, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 15970 + }, + { + "epoch": 24.030075187969924, + "grad_norm": 0.031131941825151443, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 15980 + }, + { + "epoch": 24.045112781954888, + "grad_norm": 0.017940782010555267, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 15990 + }, + { + "epoch": 24.06015037593985, + "grad_norm": 0.044517941772937775, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 16000 + }, + { + "epoch": 24.06015037593985, + "eval_cer": 0.020005739630242683, + "eval_loss": 0.08524861186742783, + "eval_runtime": 159.8819, + "eval_samples_per_second": 100.243, + "eval_steps_per_second": 0.788, + "eval_wer": 0.0697767295006484, + "step": 16000 + }, + { + "epoch": 24.075187969924812, + "grad_norm": 0.034400537610054016, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 16010 + }, + { + "epoch": 24.090225563909776, + "grad_norm": 0.056206267327070236, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 16020 + }, + { + "epoch": 24.105263157894736, + "grad_norm": 0.06723590940237045, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 16030 + }, + { + "epoch": 24.1203007518797, + "grad_norm": 0.029193704947829247, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16040 + }, + { + "epoch": 24.13533834586466, + "grad_norm": 0.024255746975541115, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 16050 + }, + { + "epoch": 24.150375939849624, + "grad_norm": 0.04077725484967232, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16060 + }, + { + "epoch": 24.165413533834588, + "grad_norm": 0.042173441499471664, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16070 + }, + { + "epoch": 24.18045112781955, + "grad_norm": 0.03520064055919647, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16080 + }, + { + "epoch": 24.195488721804512, + "grad_norm": 0.03143632411956787, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16090 + }, + { + "epoch": 24.210526315789473, + "grad_norm": 0.03446348384022713, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 16100 + }, + { + "epoch": 24.225563909774436, + "grad_norm": 0.04016732797026634, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16110 + }, + { + "epoch": 24.2406015037594, + "grad_norm": 0.040496826171875, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 16120 + }, + { + "epoch": 24.25563909774436, + "grad_norm": 0.054758574813604355, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 16130 + }, + { + "epoch": 24.270676691729324, + "grad_norm": 0.03467017039656639, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16140 + }, + { + "epoch": 24.285714285714285, + "grad_norm": 0.04085783287882805, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16150 + }, + { + "epoch": 24.30075187969925, + "grad_norm": 0.041894737631082535, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16160 + }, + { + "epoch": 24.31578947368421, + "grad_norm": 0.03293424844741821, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 16170 + }, + { + "epoch": 24.330827067669173, + "grad_norm": 0.05587819218635559, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16180 + }, + { + "epoch": 24.345864661654137, + "grad_norm": 0.04430709034204483, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16190 + }, + { + "epoch": 24.360902255639097, + "grad_norm": 0.036320075392723083, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 16200 + }, + { + "epoch": 24.37593984962406, + "grad_norm": 0.058958910405635834, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 16210 + }, + { + "epoch": 24.39097744360902, + "grad_norm": 0.03813060373067856, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16220 + }, + { + "epoch": 24.406015037593985, + "grad_norm": 0.037513598799705505, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 16230 + }, + { + "epoch": 24.42105263157895, + "grad_norm": 0.03243521973490715, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16240 + }, + { + "epoch": 24.43609022556391, + "grad_norm": 0.06187858805060387, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16250 + }, + { + "epoch": 24.451127819548873, + "grad_norm": 0.029195835813879967, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 16260 + }, + { + "epoch": 24.466165413533833, + "grad_norm": 0.029820434749126434, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 16270 + }, + { + "epoch": 24.481203007518797, + "grad_norm": 0.031871505081653595, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16280 + }, + { + "epoch": 24.49624060150376, + "grad_norm": 0.03556877747178078, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 16290 + }, + { + "epoch": 24.51127819548872, + "grad_norm": 0.03555990383028984, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16300 + }, + { + "epoch": 24.526315789473685, + "grad_norm": 0.044513195753097534, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 16310 + }, + { + "epoch": 24.541353383458645, + "grad_norm": 0.03798365592956543, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16320 + }, + { + "epoch": 24.55639097744361, + "grad_norm": 0.02286584861576557, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16330 + }, + { + "epoch": 24.571428571428573, + "grad_norm": 0.02835952863097191, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 16340 + }, + { + "epoch": 24.586466165413533, + "grad_norm": 0.04415373131632805, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16350 + }, + { + "epoch": 24.601503759398497, + "grad_norm": 0.03860164061188698, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16360 + }, + { + "epoch": 24.616541353383457, + "grad_norm": 0.030602088198065758, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16370 + }, + { + "epoch": 24.63157894736842, + "grad_norm": 0.04808523505926132, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 16380 + }, + { + "epoch": 24.646616541353385, + "grad_norm": 0.04020848497748375, + "learning_rate": 0.0005, + "loss": 0.0043, + "step": 16390 + }, + { + "epoch": 24.661654135338345, + "grad_norm": 0.047390516847372055, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16400 + }, + { + "epoch": 24.67669172932331, + "grad_norm": 0.04060389846563339, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16410 + }, + { + "epoch": 24.69172932330827, + "grad_norm": 0.042470257729291916, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16420 + }, + { + "epoch": 24.706766917293233, + "grad_norm": 0.032204803079366684, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16430 + }, + { + "epoch": 24.721804511278194, + "grad_norm": 0.07326094061136246, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16440 + }, + { + "epoch": 24.736842105263158, + "grad_norm": 0.0514272004365921, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16450 + }, + { + "epoch": 24.75187969924812, + "grad_norm": 0.038105037063360214, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16460 + }, + { + "epoch": 24.76691729323308, + "grad_norm": 0.050234124064445496, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 16470 + }, + { + "epoch": 24.781954887218046, + "grad_norm": 0.03629729524254799, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 16480 + }, + { + "epoch": 24.796992481203006, + "grad_norm": 0.03026171773672104, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 16490 + }, + { + "epoch": 24.81203007518797, + "grad_norm": 0.04203719273209572, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 16500 + }, + { + "epoch": 24.827067669172934, + "grad_norm": 0.04503991827368736, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16510 + }, + { + "epoch": 24.842105263157894, + "grad_norm": 0.04659249261021614, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 16520 + }, + { + "epoch": 24.857142857142858, + "grad_norm": 0.03899062052369118, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16530 + }, + { + "epoch": 24.872180451127818, + "grad_norm": 0.033499106764793396, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16540 + }, + { + "epoch": 24.887218045112782, + "grad_norm": 0.03691365197300911, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 16550 + }, + { + "epoch": 24.902255639097746, + "grad_norm": 0.05632985383272171, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 16560 + }, + { + "epoch": 24.917293233082706, + "grad_norm": 0.041265275329351425, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16570 + }, + { + "epoch": 24.93233082706767, + "grad_norm": 0.04158645123243332, + "learning_rate": 0.0005, + "loss": 0.0045, + "step": 16580 + }, + { + "epoch": 24.94736842105263, + "grad_norm": 0.04589154198765755, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 16590 + }, + { + "epoch": 24.962406015037594, + "grad_norm": 0.06053953990340233, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16600 + }, + { + "epoch": 24.977443609022558, + "grad_norm": 0.06355740875005722, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16610 + }, + { + "epoch": 24.992481203007518, + "grad_norm": 0.06353238224983215, + "learning_rate": 0.0005, + "loss": 0.0042, + "step": 16620 + }, + { + "epoch": 25.007518796992482, + "grad_norm": 0.028865184634923935, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 16630 + }, + { + "epoch": 25.022556390977442, + "grad_norm": 0.03024749830365181, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 16640 + }, + { + "epoch": 25.037593984962406, + "grad_norm": 0.03675810247659683, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16650 + }, + { + "epoch": 25.05263157894737, + "grad_norm": 0.056101538240909576, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16660 + }, + { + "epoch": 25.06766917293233, + "grad_norm": 0.03612141311168671, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16670 + }, + { + "epoch": 25.082706766917294, + "grad_norm": 0.05395114794373512, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16680 + }, + { + "epoch": 25.097744360902254, + "grad_norm": 0.030780382454395294, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 16690 + }, + { + "epoch": 25.112781954887218, + "grad_norm": 0.037806931883096695, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 16700 + }, + { + "epoch": 25.127819548872182, + "grad_norm": 0.03997332230210304, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 16710 + }, + { + "epoch": 25.142857142857142, + "grad_norm": 0.03846658393740654, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16720 + }, + { + "epoch": 25.157894736842106, + "grad_norm": 0.03075948916375637, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 16730 + }, + { + "epoch": 25.172932330827066, + "grad_norm": 0.03761044517159462, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 16740 + }, + { + "epoch": 25.18796992481203, + "grad_norm": 0.04369839280843735, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16750 + }, + { + "epoch": 25.203007518796994, + "grad_norm": 0.04874466359615326, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16760 + }, + { + "epoch": 25.218045112781954, + "grad_norm": 0.04877267777919769, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 16770 + }, + { + "epoch": 25.23308270676692, + "grad_norm": 0.04765179753303528, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16780 + }, + { + "epoch": 25.24812030075188, + "grad_norm": 0.04203315079212189, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 16790 + }, + { + "epoch": 25.263157894736842, + "grad_norm": 0.02651985175907612, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 16800 + }, + { + "epoch": 25.278195488721803, + "grad_norm": 0.0383647121489048, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16810 + }, + { + "epoch": 25.293233082706767, + "grad_norm": 0.05135364457964897, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16820 + }, + { + "epoch": 25.30827067669173, + "grad_norm": 0.03705194965004921, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 16830 + }, + { + "epoch": 25.32330827067669, + "grad_norm": 0.025454789400100708, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 16840 + }, + { + "epoch": 25.338345864661655, + "grad_norm": 0.046047236770391464, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 16850 + }, + { + "epoch": 25.353383458646615, + "grad_norm": 0.03577849641442299, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 16860 + }, + { + "epoch": 25.36842105263158, + "grad_norm": 0.031398359686136246, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16870 + }, + { + "epoch": 25.383458646616543, + "grad_norm": 0.037639763206243515, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 16880 + }, + { + "epoch": 25.398496240601503, + "grad_norm": 0.04259449988603592, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 16890 + }, + { + "epoch": 25.413533834586467, + "grad_norm": 0.04925589635968208, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16900 + }, + { + "epoch": 25.428571428571427, + "grad_norm": 0.0360967181622982, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16910 + }, + { + "epoch": 25.44360902255639, + "grad_norm": 0.04704995080828667, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16920 + }, + { + "epoch": 25.458646616541355, + "grad_norm": 0.03580515831708908, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 16930 + }, + { + "epoch": 25.473684210526315, + "grad_norm": 0.040450938045978546, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16940 + }, + { + "epoch": 25.48872180451128, + "grad_norm": 0.036299001425504684, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 16950 + }, + { + "epoch": 25.50375939849624, + "grad_norm": 0.030750175938010216, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16960 + }, + { + "epoch": 25.518796992481203, + "grad_norm": 0.06397616118192673, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 16970 + }, + { + "epoch": 25.533834586466167, + "grad_norm": 0.06920385360717773, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 16980 + }, + { + "epoch": 25.548872180451127, + "grad_norm": 0.0393199697136879, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 16990 + }, + { + "epoch": 25.56390977443609, + "grad_norm": 0.044527821242809296, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17000 + }, + { + "epoch": 25.56390977443609, + "eval_cer": 0.019015585408282205, + "eval_loss": 0.08520465344190598, + "eval_runtime": 162.4559, + "eval_samples_per_second": 98.654, + "eval_steps_per_second": 0.776, + "eval_wer": 0.06787385592662895, + "step": 17000 + }, + { + "epoch": 25.57894736842105, + "grad_norm": 0.04779224842786789, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17010 + }, + { + "epoch": 25.593984962406015, + "grad_norm": 0.0287516750395298, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17020 + }, + { + "epoch": 25.60902255639098, + "grad_norm": 0.06845036894083023, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 17030 + }, + { + "epoch": 25.62406015037594, + "grad_norm": 0.0304939653724432, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17040 + }, + { + "epoch": 25.639097744360903, + "grad_norm": 0.03194738179445267, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17050 + }, + { + "epoch": 25.654135338345863, + "grad_norm": 0.029131758958101273, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17060 + }, + { + "epoch": 25.669172932330827, + "grad_norm": 0.0656234622001648, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17070 + }, + { + "epoch": 25.68421052631579, + "grad_norm": 0.03815919905900955, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17080 + }, + { + "epoch": 25.69924812030075, + "grad_norm": 0.0376497320830822, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 17090 + }, + { + "epoch": 25.714285714285715, + "grad_norm": 0.036683470010757446, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17100 + }, + { + "epoch": 25.729323308270676, + "grad_norm": 0.04406420886516571, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17110 + }, + { + "epoch": 25.74436090225564, + "grad_norm": 0.04158684238791466, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17120 + }, + { + "epoch": 25.7593984962406, + "grad_norm": 0.08397620171308517, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 17130 + }, + { + "epoch": 25.774436090225564, + "grad_norm": 0.03754282370209694, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17140 + }, + { + "epoch": 25.789473684210527, + "grad_norm": 0.03342588618397713, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 17150 + }, + { + "epoch": 25.804511278195488, + "grad_norm": 0.051246967166662216, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17160 + }, + { + "epoch": 25.81954887218045, + "grad_norm": 0.0333494134247303, + "learning_rate": 0.0005, + "loss": 0.004, + "step": 17170 + }, + { + "epoch": 25.834586466165412, + "grad_norm": 0.03742311894893646, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 17180 + }, + { + "epoch": 25.849624060150376, + "grad_norm": 0.04523573815822601, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 17190 + }, + { + "epoch": 25.86466165413534, + "grad_norm": 0.0417187437415123, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17200 + }, + { + "epoch": 25.8796992481203, + "grad_norm": 0.029240217059850693, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17210 + }, + { + "epoch": 25.894736842105264, + "grad_norm": 0.03455571457743645, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17220 + }, + { + "epoch": 25.909774436090224, + "grad_norm": 0.022237218916416168, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 17230 + }, + { + "epoch": 25.924812030075188, + "grad_norm": 0.03676801547408104, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17240 + }, + { + "epoch": 25.93984962406015, + "grad_norm": 0.038016606122255325, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17250 + }, + { + "epoch": 25.954887218045112, + "grad_norm": 0.0325593501329422, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 17260 + }, + { + "epoch": 25.969924812030076, + "grad_norm": 0.04499173164367676, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17270 + }, + { + "epoch": 25.984962406015036, + "grad_norm": 0.03318900987505913, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17280 + }, + { + "epoch": 26.0, + "grad_norm": 0.03884175419807434, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 17290 + }, + { + "epoch": 26.015037593984964, + "grad_norm": 0.03231540322303772, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 17300 + }, + { + "epoch": 26.030075187969924, + "grad_norm": 0.026901941746473312, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 17310 + }, + { + "epoch": 26.045112781954888, + "grad_norm": 0.03841046243906021, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 17320 + }, + { + "epoch": 26.06015037593985, + "grad_norm": 0.031196730211377144, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 17330 + }, + { + "epoch": 26.075187969924812, + "grad_norm": 0.02794775739312172, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 17340 + }, + { + "epoch": 26.090225563909776, + "grad_norm": 0.03437412530183792, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 17350 + }, + { + "epoch": 26.105263157894736, + "grad_norm": 0.034914225339889526, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 17360 + }, + { + "epoch": 26.1203007518797, + "grad_norm": 0.04672201722860336, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17370 + }, + { + "epoch": 26.13533834586466, + "grad_norm": 0.037713512778282166, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 17380 + }, + { + "epoch": 26.150375939849624, + "grad_norm": 0.044108759611845016, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 17390 + }, + { + "epoch": 26.165413533834588, + "grad_norm": 0.03144936263561249, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 17400 + }, + { + "epoch": 26.18045112781955, + "grad_norm": 0.03708245977759361, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 17410 + }, + { + "epoch": 26.195488721804512, + "grad_norm": 0.035425614565610886, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 17420 + }, + { + "epoch": 26.210526315789473, + "grad_norm": 0.036712341010570526, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 17430 + }, + { + "epoch": 26.225563909774436, + "grad_norm": 0.04577183350920677, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17440 + }, + { + "epoch": 26.2406015037594, + "grad_norm": 0.05120034143328667, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17450 + }, + { + "epoch": 26.25563909774436, + "grad_norm": 0.0431804396212101, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17460 + }, + { + "epoch": 26.270676691729324, + "grad_norm": 0.036451853811740875, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 17470 + }, + { + "epoch": 26.285714285714285, + "grad_norm": 0.06433060765266418, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17480 + }, + { + "epoch": 26.30075187969925, + "grad_norm": 0.02980835549533367, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17490 + }, + { + "epoch": 26.31578947368421, + "grad_norm": 0.025401024147868156, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17500 + }, + { + "epoch": 26.330827067669173, + "grad_norm": 0.05200992152094841, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17510 + }, + { + "epoch": 26.345864661654137, + "grad_norm": 0.036107685416936874, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17520 + }, + { + "epoch": 26.360902255639097, + "grad_norm": 0.02595752477645874, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17530 + }, + { + "epoch": 26.37593984962406, + "grad_norm": 0.03037913143634796, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17540 + }, + { + "epoch": 26.39097744360902, + "grad_norm": 0.05118720978498459, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17550 + }, + { + "epoch": 26.406015037593985, + "grad_norm": 0.026243647560477257, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17560 + }, + { + "epoch": 26.42105263157895, + "grad_norm": 0.03880322352051735, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17570 + }, + { + "epoch": 26.43609022556391, + "grad_norm": 0.025365188717842102, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17580 + }, + { + "epoch": 26.451127819548873, + "grad_norm": 0.038137286901474, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17590 + }, + { + "epoch": 26.466165413533833, + "grad_norm": 0.03465953841805458, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17600 + }, + { + "epoch": 26.481203007518797, + "grad_norm": 0.04526372626423836, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17610 + }, + { + "epoch": 26.49624060150376, + "grad_norm": 0.02651445008814335, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17620 + }, + { + "epoch": 26.51127819548872, + "grad_norm": 0.04033152759075165, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17630 + }, + { + "epoch": 26.526315789473685, + "grad_norm": 0.031117983162403107, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 17640 + }, + { + "epoch": 26.541353383458645, + "grad_norm": 0.03773597627878189, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17650 + }, + { + "epoch": 26.55639097744361, + "grad_norm": 0.05135533958673477, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17660 + }, + { + "epoch": 26.571428571428573, + "grad_norm": 0.03314359486103058, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 17670 + }, + { + "epoch": 26.586466165413533, + "grad_norm": 0.04010153189301491, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17680 + }, + { + "epoch": 26.601503759398497, + "grad_norm": 0.03848304599523544, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17690 + }, + { + "epoch": 26.616541353383457, + "grad_norm": 0.0327029712498188, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 17700 + }, + { + "epoch": 26.63157894736842, + "grad_norm": 0.041056521236896515, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17710 + }, + { + "epoch": 26.646616541353385, + "grad_norm": 0.06059001386165619, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17720 + }, + { + "epoch": 26.661654135338345, + "grad_norm": 0.03288688138127327, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 17730 + }, + { + "epoch": 26.67669172932331, + "grad_norm": 0.04945435747504234, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 17740 + }, + { + "epoch": 26.69172932330827, + "grad_norm": 0.042480625212192535, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17750 + }, + { + "epoch": 26.706766917293233, + "grad_norm": 0.03110503777861595, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 17760 + }, + { + "epoch": 26.721804511278194, + "grad_norm": 0.03717326000332832, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17770 + }, + { + "epoch": 26.736842105263158, + "grad_norm": 0.034811388701200485, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17780 + }, + { + "epoch": 26.75187969924812, + "grad_norm": 0.0349460132420063, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 17790 + }, + { + "epoch": 26.76691729323308, + "grad_norm": 0.043027158826589584, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17800 + }, + { + "epoch": 26.781954887218046, + "grad_norm": 0.04734886437654495, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17810 + }, + { + "epoch": 26.796992481203006, + "grad_norm": 0.032076865434646606, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17820 + }, + { + "epoch": 26.81203007518797, + "grad_norm": 0.038637202233076096, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17830 + }, + { + "epoch": 26.827067669172934, + "grad_norm": 0.03524789214134216, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17840 + }, + { + "epoch": 26.842105263157894, + "grad_norm": 0.04380466043949127, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 17850 + }, + { + "epoch": 26.857142857142858, + "grad_norm": 0.04988179728388786, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 17860 + }, + { + "epoch": 26.872180451127818, + "grad_norm": 0.03055410645902157, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 17870 + }, + { + "epoch": 26.887218045112782, + "grad_norm": 0.04004744812846184, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 17880 + }, + { + "epoch": 26.902255639097746, + "grad_norm": 0.03533731400966644, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17890 + }, + { + "epoch": 26.917293233082706, + "grad_norm": 0.053682029247283936, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 17900 + }, + { + "epoch": 26.93233082706767, + "grad_norm": 0.03692735731601715, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 17910 + }, + { + "epoch": 26.94736842105263, + "grad_norm": 0.062043171375989914, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 17920 + }, + { + "epoch": 26.962406015037594, + "grad_norm": 0.03803117200732231, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 17930 + }, + { + "epoch": 26.977443609022558, + "grad_norm": 0.03888007998466492, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 17940 + }, + { + "epoch": 26.992481203007518, + "grad_norm": 0.042125336825847626, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 17950 + }, + { + "epoch": 27.007518796992482, + "grad_norm": 0.06511995196342468, + "learning_rate": 0.0005, + "loss": 0.0039, + "step": 17960 + }, + { + "epoch": 27.022556390977442, + "grad_norm": 0.031230000779032707, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 17970 + }, + { + "epoch": 27.037593984962406, + "grad_norm": 0.041249122470617294, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 17980 + }, + { + "epoch": 27.05263157894737, + "grad_norm": 0.03678766265511513, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 17990 + }, + { + "epoch": 27.06766917293233, + "grad_norm": 0.028801018372178078, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18000 + }, + { + "epoch": 27.06766917293233, + "eval_cer": 0.019592268636456987, + "eval_loss": 0.08845138549804688, + "eval_runtime": 161.2133, + "eval_samples_per_second": 99.415, + "eval_steps_per_second": 0.782, + "eval_wer": 0.07042511605179574, + "step": 18000 + }, + { + "epoch": 27.082706766917294, + "grad_norm": 0.021794892847537994, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18010 + }, + { + "epoch": 27.097744360902254, + "grad_norm": 0.03247411549091339, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18020 + }, + { + "epoch": 27.112781954887218, + "grad_norm": 0.05231059715151787, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18030 + }, + { + "epoch": 27.127819548872182, + "grad_norm": 0.03336777910590172, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18040 + }, + { + "epoch": 27.142857142857142, + "grad_norm": 0.04830486327409744, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18050 + }, + { + "epoch": 27.157894736842106, + "grad_norm": 0.036319270730018616, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18060 + }, + { + "epoch": 27.172932330827066, + "grad_norm": 0.0348498709499836, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 18070 + }, + { + "epoch": 27.18796992481203, + "grad_norm": 0.03911750391125679, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18080 + }, + { + "epoch": 27.203007518796994, + "grad_norm": 0.03706942871212959, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 18090 + }, + { + "epoch": 27.218045112781954, + "grad_norm": 0.030709289014339447, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18100 + }, + { + "epoch": 27.23308270676692, + "grad_norm": 0.043648768216371536, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18110 + }, + { + "epoch": 27.24812030075188, + "grad_norm": 0.03740646690130234, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18120 + }, + { + "epoch": 27.263157894736842, + "grad_norm": 0.029137492179870605, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18130 + }, + { + "epoch": 27.278195488721803, + "grad_norm": 0.04714982211589813, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18140 + }, + { + "epoch": 27.293233082706767, + "grad_norm": 0.052695658057928085, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18150 + }, + { + "epoch": 27.30827067669173, + "grad_norm": 0.0393044650554657, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18160 + }, + { + "epoch": 27.32330827067669, + "grad_norm": 0.03941282257437706, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18170 + }, + { + "epoch": 27.338345864661655, + "grad_norm": 0.03247359022498131, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18180 + }, + { + "epoch": 27.353383458646615, + "grad_norm": 0.043848223984241486, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 18190 + }, + { + "epoch": 27.36842105263158, + "grad_norm": 0.04029109328985214, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18200 + }, + { + "epoch": 27.383458646616543, + "grad_norm": 0.0185089111328125, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18210 + }, + { + "epoch": 27.398496240601503, + "grad_norm": 0.032335154712200165, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 18220 + }, + { + "epoch": 27.413533834586467, + "grad_norm": 0.030238516628742218, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18230 + }, + { + "epoch": 27.428571428571427, + "grad_norm": 0.03891552984714508, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18240 + }, + { + "epoch": 27.44360902255639, + "grad_norm": 0.03493286296725273, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18250 + }, + { + "epoch": 27.458646616541355, + "grad_norm": 0.02859729714691639, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18260 + }, + { + "epoch": 27.473684210526315, + "grad_norm": 0.029383642598986626, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18270 + }, + { + "epoch": 27.48872180451128, + "grad_norm": 0.036221183836460114, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 18280 + }, + { + "epoch": 27.50375939849624, + "grad_norm": 0.036437541246414185, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18290 + }, + { + "epoch": 27.518796992481203, + "grad_norm": 0.031798187643289566, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18300 + }, + { + "epoch": 27.533834586466167, + "grad_norm": 0.04122854396700859, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18310 + }, + { + "epoch": 27.548872180451127, + "grad_norm": 0.030801068991422653, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18320 + }, + { + "epoch": 27.56390977443609, + "grad_norm": 0.044105302542448044, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18330 + }, + { + "epoch": 27.57894736842105, + "grad_norm": 0.03397545590996742, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18340 + }, + { + "epoch": 27.593984962406015, + "grad_norm": 0.03347723186016083, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18350 + }, + { + "epoch": 27.60902255639098, + "grad_norm": 0.027925802394747734, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18360 + }, + { + "epoch": 27.62406015037594, + "grad_norm": 0.04396814480423927, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18370 + }, + { + "epoch": 27.639097744360903, + "grad_norm": 0.03711283951997757, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18380 + }, + { + "epoch": 27.654135338345863, + "grad_norm": 0.034581080079078674, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 18390 + }, + { + "epoch": 27.669172932330827, + "grad_norm": 0.05024787783622742, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18400 + }, + { + "epoch": 27.68421052631579, + "grad_norm": 0.04361278563737869, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 18410 + }, + { + "epoch": 27.69924812030075, + "grad_norm": 0.04109452664852142, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18420 + }, + { + "epoch": 27.714285714285715, + "grad_norm": 0.04303286224603653, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 18430 + }, + { + "epoch": 27.729323308270676, + "grad_norm": 0.04334039241075516, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 18440 + }, + { + "epoch": 27.74436090225564, + "grad_norm": 0.03496752679347992, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 18450 + }, + { + "epoch": 27.7593984962406, + "grad_norm": 0.08848168700933456, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 18460 + }, + { + "epoch": 27.774436090225564, + "grad_norm": 0.0357455275952816, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 18470 + }, + { + "epoch": 27.789473684210527, + "grad_norm": 0.03218315541744232, + "learning_rate": 0.0005, + "loss": 0.0041, + "step": 18480 + }, + { + "epoch": 27.804511278195488, + "grad_norm": 0.024740636348724365, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18490 + }, + { + "epoch": 27.81954887218045, + "grad_norm": 0.03971486538648605, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 18500 + }, + { + "epoch": 27.834586466165412, + "grad_norm": 0.0528874397277832, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 18510 + }, + { + "epoch": 27.849624060150376, + "grad_norm": 0.027518296614289284, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 18520 + }, + { + "epoch": 27.86466165413534, + "grad_norm": 0.03509645536541939, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 18530 + }, + { + "epoch": 27.8796992481203, + "grad_norm": 0.03134102001786232, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18540 + }, + { + "epoch": 27.894736842105264, + "grad_norm": 0.049625035375356674, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18550 + }, + { + "epoch": 27.909774436090224, + "grad_norm": 0.03801269084215164, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 18560 + }, + { + "epoch": 27.924812030075188, + "grad_norm": 0.04829275235533714, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 18570 + }, + { + "epoch": 27.93984962406015, + "grad_norm": 0.06010720506310463, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 18580 + }, + { + "epoch": 27.954887218045112, + "grad_norm": 0.042644064873456955, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 18590 + }, + { + "epoch": 27.969924812030076, + "grad_norm": 0.05876699835062027, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18600 + }, + { + "epoch": 27.984962406015036, + "grad_norm": 0.032208703458309174, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 18610 + }, + { + "epoch": 28.0, + "grad_norm": 0.0428246445953846, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18620 + }, + { + "epoch": 28.015037593984964, + "grad_norm": 0.039440639317035675, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18630 + }, + { + "epoch": 28.030075187969924, + "grad_norm": 0.028335612267255783, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 18640 + }, + { + "epoch": 28.045112781954888, + "grad_norm": 0.021635161712765694, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 18650 + }, + { + "epoch": 28.06015037593985, + "grad_norm": 0.03994419425725937, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18660 + }, + { + "epoch": 28.075187969924812, + "grad_norm": 0.05112672597169876, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18670 + }, + { + "epoch": 28.090225563909776, + "grad_norm": 0.03884351998567581, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18680 + }, + { + "epoch": 28.105263157894736, + "grad_norm": 0.03763177618384361, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 18690 + }, + { + "epoch": 28.1203007518797, + "grad_norm": 0.04131988808512688, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 18700 + }, + { + "epoch": 28.13533834586466, + "grad_norm": 0.04687080159783363, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18710 + }, + { + "epoch": 28.150375939849624, + "grad_norm": 0.02924514375627041, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18720 + }, + { + "epoch": 28.165413533834588, + "grad_norm": 0.04289636015892029, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18730 + }, + { + "epoch": 28.18045112781955, + "grad_norm": 0.041429225355386734, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18740 + }, + { + "epoch": 28.195488721804512, + "grad_norm": 0.03454408049583435, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18750 + }, + { + "epoch": 28.210526315789473, + "grad_norm": 0.030546654015779495, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 18760 + }, + { + "epoch": 28.225563909774436, + "grad_norm": 0.045081790536642075, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18770 + }, + { + "epoch": 28.2406015037594, + "grad_norm": 0.036060530692338943, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18780 + }, + { + "epoch": 28.25563909774436, + "grad_norm": 0.05681438371539116, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18790 + }, + { + "epoch": 28.270676691729324, + "grad_norm": 0.033227480947971344, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18800 + }, + { + "epoch": 28.285714285714285, + "grad_norm": 0.03013618476688862, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18810 + }, + { + "epoch": 28.30075187969925, + "grad_norm": 0.043254077434539795, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 18820 + }, + { + "epoch": 28.31578947368421, + "grad_norm": 0.028685178607702255, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18830 + }, + { + "epoch": 28.330827067669173, + "grad_norm": 0.026451099663972855, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18840 + }, + { + "epoch": 28.345864661654137, + "grad_norm": 0.030207518488168716, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18850 + }, + { + "epoch": 28.360902255639097, + "grad_norm": 0.037875302135944366, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18860 + }, + { + "epoch": 28.37593984962406, + "grad_norm": 0.031988803297281265, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 18870 + }, + { + "epoch": 28.39097744360902, + "grad_norm": 0.03780095651745796, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 18880 + }, + { + "epoch": 28.406015037593985, + "grad_norm": 0.03800348564982414, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18890 + }, + { + "epoch": 28.42105263157895, + "grad_norm": 0.041992221027612686, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18900 + }, + { + "epoch": 28.43609022556391, + "grad_norm": 0.03876485303044319, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18910 + }, + { + "epoch": 28.451127819548873, + "grad_norm": 0.05463851988315582, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18920 + }, + { + "epoch": 28.466165413533833, + "grad_norm": 0.028002101927995682, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 18930 + }, + { + "epoch": 28.481203007518797, + "grad_norm": 0.027172712609171867, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 18940 + }, + { + "epoch": 28.49624060150376, + "grad_norm": 0.040023159235715866, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 18950 + }, + { + "epoch": 28.51127819548872, + "grad_norm": 0.03482840955257416, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18960 + }, + { + "epoch": 28.526315789473685, + "grad_norm": 0.02487008459866047, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 18970 + }, + { + "epoch": 28.541353383458645, + "grad_norm": 0.03150250017642975, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 18980 + }, + { + "epoch": 28.55639097744361, + "grad_norm": 0.03910740464925766, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 18990 + }, + { + "epoch": 28.571428571428573, + "grad_norm": 0.040712758898735046, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19000 + }, + { + "epoch": 28.571428571428573, + "eval_cer": 0.020315162824605333, + "eval_loss": 0.08566422015428543, + "eval_runtime": 162.3898, + "eval_samples_per_second": 98.695, + "eval_steps_per_second": 0.776, + "eval_wer": 0.06981431712680186, + "step": 19000 + }, + { + "epoch": 28.586466165413533, + "grad_norm": 0.050952911376953125, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 19010 + }, + { + "epoch": 28.601503759398497, + "grad_norm": 0.043291185051202774, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19020 + }, + { + "epoch": 28.616541353383457, + "grad_norm": 0.04345792159438133, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 19030 + }, + { + "epoch": 28.63157894736842, + "grad_norm": 0.031741295009851456, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19040 + }, + { + "epoch": 28.646616541353385, + "grad_norm": 0.045271411538124084, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19050 + }, + { + "epoch": 28.661654135338345, + "grad_norm": 0.038084592670202255, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 19060 + }, + { + "epoch": 28.67669172932331, + "grad_norm": 0.04593008756637573, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 19070 + }, + { + "epoch": 28.69172932330827, + "grad_norm": 0.04494618624448776, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19080 + }, + { + "epoch": 28.706766917293233, + "grad_norm": 0.04930887743830681, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19090 + }, + { + "epoch": 28.721804511278194, + "grad_norm": 0.03185184299945831, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19100 + }, + { + "epoch": 28.736842105263158, + "grad_norm": 0.0306316576898098, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19110 + }, + { + "epoch": 28.75187969924812, + "grad_norm": 0.03257163614034653, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19120 + }, + { + "epoch": 28.76691729323308, + "grad_norm": 0.0249934084713459, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19130 + }, + { + "epoch": 28.781954887218046, + "grad_norm": 0.027818778529763222, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19140 + }, + { + "epoch": 28.796992481203006, + "grad_norm": 0.0317297987639904, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19150 + }, + { + "epoch": 28.81203007518797, + "grad_norm": 0.03958116099238396, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19160 + }, + { + "epoch": 28.827067669172934, + "grad_norm": 0.03359660133719444, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19170 + }, + { + "epoch": 28.842105263157894, + "grad_norm": 0.05302347615361214, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 19180 + }, + { + "epoch": 28.857142857142858, + "grad_norm": 0.026908375322818756, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19190 + }, + { + "epoch": 28.872180451127818, + "grad_norm": 0.049994513392448425, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19200 + }, + { + "epoch": 28.887218045112782, + "grad_norm": 0.04454610496759415, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 19210 + }, + { + "epoch": 28.902255639097746, + "grad_norm": 0.03436105698347092, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19220 + }, + { + "epoch": 28.917293233082706, + "grad_norm": 0.04194503277540207, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19230 + }, + { + "epoch": 28.93233082706767, + "grad_norm": 0.035593245178461075, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 19240 + }, + { + "epoch": 28.94736842105263, + "grad_norm": 0.02727363258600235, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 19250 + }, + { + "epoch": 28.962406015037594, + "grad_norm": 0.03494620323181152, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19260 + }, + { + "epoch": 28.977443609022558, + "grad_norm": 0.03637601435184479, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19270 + }, + { + "epoch": 28.992481203007518, + "grad_norm": 0.04141705110669136, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19280 + }, + { + "epoch": 29.007518796992482, + "grad_norm": 0.032516639679670334, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19290 + }, + { + "epoch": 29.022556390977442, + "grad_norm": 0.025468016043305397, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 19300 + }, + { + "epoch": 29.037593984962406, + "grad_norm": 0.025055214762687683, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 19310 + }, + { + "epoch": 29.05263157894737, + "grad_norm": 0.042539458721876144, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 19320 + }, + { + "epoch": 29.06766917293233, + "grad_norm": 0.033082231879234314, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 19330 + }, + { + "epoch": 29.082706766917294, + "grad_norm": 0.030265340581536293, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19340 + }, + { + "epoch": 29.097744360902254, + "grad_norm": 0.047304246574640274, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19350 + }, + { + "epoch": 29.112781954887218, + "grad_norm": 0.041672371327877045, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 19360 + }, + { + "epoch": 29.127819548872182, + "grad_norm": 0.030016224831342697, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 19370 + }, + { + "epoch": 29.142857142857142, + "grad_norm": 0.051045503467321396, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19380 + }, + { + "epoch": 29.157894736842106, + "grad_norm": 0.027139203622937202, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 19390 + }, + { + "epoch": 29.172932330827066, + "grad_norm": 0.04479178041219711, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19400 + }, + { + "epoch": 29.18796992481203, + "grad_norm": 0.041574664413928986, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19410 + }, + { + "epoch": 29.203007518796994, + "grad_norm": 0.036204490810632706, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19420 + }, + { + "epoch": 29.218045112781954, + "grad_norm": 0.03129658102989197, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 19430 + }, + { + "epoch": 29.23308270676692, + "grad_norm": 0.02882031537592411, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19440 + }, + { + "epoch": 29.24812030075188, + "grad_norm": 0.032281313091516495, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19450 + }, + { + "epoch": 29.263157894736842, + "grad_norm": 0.04641575738787651, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19460 + }, + { + "epoch": 29.278195488721803, + "grad_norm": 0.049606479704380035, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 19470 + }, + { + "epoch": 29.293233082706767, + "grad_norm": 0.03462973237037659, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19480 + }, + { + "epoch": 29.30827067669173, + "grad_norm": 0.036535680294036865, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19490 + }, + { + "epoch": 29.32330827067669, + "grad_norm": 0.03790770098567009, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19500 + }, + { + "epoch": 29.338345864661655, + "grad_norm": 0.04262432083487511, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19510 + }, + { + "epoch": 29.353383458646615, + "grad_norm": 0.03205105662345886, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 19520 + }, + { + "epoch": 29.36842105263158, + "grad_norm": 0.03708963468670845, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19530 + }, + { + "epoch": 29.383458646616543, + "grad_norm": 0.046063121408224106, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19540 + }, + { + "epoch": 29.398496240601503, + "grad_norm": 0.040626950562000275, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 19550 + }, + { + "epoch": 29.413533834586467, + "grad_norm": 0.029632287099957466, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19560 + }, + { + "epoch": 29.428571428571427, + "grad_norm": 0.04287538304924965, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19570 + }, + { + "epoch": 29.44360902255639, + "grad_norm": 0.03313089907169342, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 19580 + }, + { + "epoch": 29.458646616541355, + "grad_norm": 0.027183614671230316, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 19590 + }, + { + "epoch": 29.473684210526315, + "grad_norm": 0.03642988204956055, + "learning_rate": 0.0005, + "loss": 0.0037, + "step": 19600 + }, + { + "epoch": 29.48872180451128, + "grad_norm": 0.03818807378411293, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19610 + }, + { + "epoch": 29.50375939849624, + "grad_norm": 0.06420944631099701, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19620 + }, + { + "epoch": 29.518796992481203, + "grad_norm": 0.03600897639989853, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19630 + }, + { + "epoch": 29.533834586466167, + "grad_norm": 0.028436269611120224, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19640 + }, + { + "epoch": 29.548872180451127, + "grad_norm": 0.04273846745491028, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19650 + }, + { + "epoch": 29.56390977443609, + "grad_norm": 0.04267692565917969, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19660 + }, + { + "epoch": 29.57894736842105, + "grad_norm": 0.05707879364490509, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 19670 + }, + { + "epoch": 29.593984962406015, + "grad_norm": 0.040246862918138504, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19680 + }, + { + "epoch": 29.60902255639098, + "grad_norm": 0.03213886916637421, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 19690 + }, + { + "epoch": 29.62406015037594, + "grad_norm": 0.048189494758844376, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19700 + }, + { + "epoch": 29.639097744360903, + "grad_norm": 0.03790193051099777, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 19710 + }, + { + "epoch": 29.654135338345863, + "grad_norm": 0.03298810124397278, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19720 + }, + { + "epoch": 29.669172932330827, + "grad_norm": 0.04650110378861427, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19730 + }, + { + "epoch": 29.68421052631579, + "grad_norm": 0.03043610230088234, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19740 + }, + { + "epoch": 29.69924812030075, + "grad_norm": 0.029208796098828316, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19750 + }, + { + "epoch": 29.714285714285715, + "grad_norm": 0.028749484568834305, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19760 + }, + { + "epoch": 29.729323308270676, + "grad_norm": 0.021796472370624542, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19770 + }, + { + "epoch": 29.74436090225564, + "grad_norm": 0.03557973727583885, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19780 + }, + { + "epoch": 29.7593984962406, + "grad_norm": 0.03215000778436661, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 19790 + }, + { + "epoch": 29.774436090225564, + "grad_norm": 0.029668748378753662, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19800 + }, + { + "epoch": 29.789473684210527, + "grad_norm": 0.03379977121949196, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19810 + }, + { + "epoch": 29.804511278195488, + "grad_norm": 0.036043617874383926, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 19820 + }, + { + "epoch": 29.81954887218045, + "grad_norm": 0.03143557906150818, + "learning_rate": 0.0005, + "loss": 0.0036, + "step": 19830 + }, + { + "epoch": 29.834586466165412, + "grad_norm": 0.02848399057984352, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 19840 + }, + { + "epoch": 29.849624060150376, + "grad_norm": 0.03267832472920418, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 19850 + }, + { + "epoch": 29.86466165413534, + "grad_norm": 0.026367779821157455, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 19860 + }, + { + "epoch": 29.8796992481203, + "grad_norm": 0.032531727105379105, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 19870 + }, + { + "epoch": 29.894736842105264, + "grad_norm": 0.05030421167612076, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19880 + }, + { + "epoch": 29.909774436090224, + "grad_norm": 0.05682966858148575, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 19890 + }, + { + "epoch": 29.924812030075188, + "grad_norm": 0.059888169169425964, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 19900 + }, + { + "epoch": 29.93984962406015, + "grad_norm": 0.03770558908581734, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19910 + }, + { + "epoch": 29.954887218045112, + "grad_norm": 0.038983915001153946, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 19920 + }, + { + "epoch": 29.969924812030076, + "grad_norm": 0.024569351226091385, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 19930 + }, + { + "epoch": 29.984962406015036, + "grad_norm": 0.03012099303305149, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 19940 + }, + { + "epoch": 30.0, + "grad_norm": 0.05479085072875023, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 19950 + }, + { + "epoch": 30.015037593984964, + "grad_norm": 0.03115060180425644, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19960 + }, + { + "epoch": 30.030075187969924, + "grad_norm": 0.03997494652867317, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 19970 + }, + { + "epoch": 30.045112781954888, + "grad_norm": 0.033401817083358765, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 19980 + }, + { + "epoch": 30.06015037593985, + "grad_norm": 0.029517218470573425, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 19990 + }, + { + "epoch": 30.075187969924812, + "grad_norm": 0.032621242105960846, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20000 + }, + { + "epoch": 30.075187969924812, + "eval_cer": 0.01931752804190202, + "eval_loss": 0.08742145448923111, + "eval_runtime": 161.3864, + "eval_samples_per_second": 99.308, + "eval_steps_per_second": 0.781, + "eval_wer": 0.06784096675374467, + "step": 20000 + }, + { + "epoch": 30.090225563909776, + "grad_norm": 0.045351140201091766, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20010 + }, + { + "epoch": 30.105263157894736, + "grad_norm": 0.03234616294503212, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 20020 + }, + { + "epoch": 30.1203007518797, + "grad_norm": 0.039552994072437286, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20030 + }, + { + "epoch": 30.13533834586466, + "grad_norm": 0.020054170861840248, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20040 + }, + { + "epoch": 30.150375939849624, + "grad_norm": 0.04938104376196861, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20050 + }, + { + "epoch": 30.165413533834588, + "grad_norm": 0.03934120759367943, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20060 + }, + { + "epoch": 30.18045112781955, + "grad_norm": 0.027889039367437363, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20070 + }, + { + "epoch": 30.195488721804512, + "grad_norm": 0.020225761458277702, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20080 + }, + { + "epoch": 30.210526315789473, + "grad_norm": 0.03410186246037483, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 20090 + }, + { + "epoch": 30.225563909774436, + "grad_norm": 0.021709734573960304, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20100 + }, + { + "epoch": 30.2406015037594, + "grad_norm": 0.033819906413555145, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20110 + }, + { + "epoch": 30.25563909774436, + "grad_norm": 0.027473323047161102, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20120 + }, + { + "epoch": 30.270676691729324, + "grad_norm": 0.021676164120435715, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20130 + }, + { + "epoch": 30.285714285714285, + "grad_norm": 0.03573176637291908, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20140 + }, + { + "epoch": 30.30075187969925, + "grad_norm": 0.04019770398736, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20150 + }, + { + "epoch": 30.31578947368421, + "grad_norm": 0.026702648028731346, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20160 + }, + { + "epoch": 30.330827067669173, + "grad_norm": 0.027424154803156853, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20170 + }, + { + "epoch": 30.345864661654137, + "grad_norm": 0.031308289617300034, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20180 + }, + { + "epoch": 30.360902255639097, + "grad_norm": 0.03551609069108963, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20190 + }, + { + "epoch": 30.37593984962406, + "grad_norm": 0.04532497748732567, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20200 + }, + { + "epoch": 30.39097744360902, + "grad_norm": 0.031929053366184235, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 20210 + }, + { + "epoch": 30.406015037593985, + "grad_norm": 0.04581012576818466, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20220 + }, + { + "epoch": 30.42105263157895, + "grad_norm": 0.032717298716306686, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 20230 + }, + { + "epoch": 30.43609022556391, + "grad_norm": 0.028632590547204018, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20240 + }, + { + "epoch": 30.451127819548873, + "grad_norm": 0.02960212714970112, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20250 + }, + { + "epoch": 30.466165413533833, + "grad_norm": 0.026230327785015106, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20260 + }, + { + "epoch": 30.481203007518797, + "grad_norm": 0.031774912029504776, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20270 + }, + { + "epoch": 30.49624060150376, + "grad_norm": 0.04631762206554413, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20280 + }, + { + "epoch": 30.51127819548872, + "grad_norm": 0.041037414222955704, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20290 + }, + { + "epoch": 30.526315789473685, + "grad_norm": 0.02734365500509739, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 20300 + }, + { + "epoch": 30.541353383458645, + "grad_norm": 0.040381476283073425, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20310 + }, + { + "epoch": 30.55639097744361, + "grad_norm": 0.043123260140419006, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20320 + }, + { + "epoch": 30.571428571428573, + "grad_norm": 0.027452118694782257, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20330 + }, + { + "epoch": 30.586466165413533, + "grad_norm": 0.03622806817293167, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20340 + }, + { + "epoch": 30.601503759398497, + "grad_norm": 0.03801897168159485, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20350 + }, + { + "epoch": 30.616541353383457, + "grad_norm": 0.04272555932402611, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20360 + }, + { + "epoch": 30.63157894736842, + "grad_norm": 0.02489612065255642, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 20370 + }, + { + "epoch": 30.646616541353385, + "grad_norm": 0.03694657236337662, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20380 + }, + { + "epoch": 30.661654135338345, + "grad_norm": 0.02831796370446682, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20390 + }, + { + "epoch": 30.67669172932331, + "grad_norm": 0.036624811589717865, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20400 + }, + { + "epoch": 30.69172932330827, + "grad_norm": 0.02935672551393509, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 20410 + }, + { + "epoch": 30.706766917293233, + "grad_norm": 0.04537936672568321, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 20420 + }, + { + "epoch": 30.721804511278194, + "grad_norm": 0.025472817942500114, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 20430 + }, + { + "epoch": 30.736842105263158, + "grad_norm": 0.05328234285116196, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20440 + }, + { + "epoch": 30.75187969924812, + "grad_norm": 0.03392201289534569, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20450 + }, + { + "epoch": 30.76691729323308, + "grad_norm": 0.030926048755645752, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20460 + }, + { + "epoch": 30.781954887218046, + "grad_norm": 0.03927863761782646, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20470 + }, + { + "epoch": 30.796992481203006, + "grad_norm": 0.022855330258607864, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20480 + }, + { + "epoch": 30.81203007518797, + "grad_norm": 0.0351700596511364, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 20490 + }, + { + "epoch": 30.827067669172934, + "grad_norm": 0.038094621151685715, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 20500 + }, + { + "epoch": 30.842105263157894, + "grad_norm": 0.04867769032716751, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20510 + }, + { + "epoch": 30.857142857142858, + "grad_norm": 0.023385141044855118, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20520 + }, + { + "epoch": 30.872180451127818, + "grad_norm": 0.04016038402915001, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20530 + }, + { + "epoch": 30.887218045112782, + "grad_norm": 0.03656616061925888, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20540 + }, + { + "epoch": 30.902255639097746, + "grad_norm": 0.038978129625320435, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20550 + }, + { + "epoch": 30.917293233082706, + "grad_norm": 0.038148149847984314, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20560 + }, + { + "epoch": 30.93233082706767, + "grad_norm": 0.02454294264316559, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20570 + }, + { + "epoch": 30.94736842105263, + "grad_norm": 0.02689015492796898, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20580 + }, + { + "epoch": 30.962406015037594, + "grad_norm": 0.03632423281669617, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20590 + }, + { + "epoch": 30.977443609022558, + "grad_norm": 0.031994715332984924, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 20600 + }, + { + "epoch": 30.992481203007518, + "grad_norm": 0.03108692727982998, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20610 + }, + { + "epoch": 31.007518796992482, + "grad_norm": 0.02618470974266529, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 20620 + }, + { + "epoch": 31.022556390977442, + "grad_norm": 0.019812824204564095, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 20630 + }, + { + "epoch": 31.037593984962406, + "grad_norm": 0.03418208286166191, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 20640 + }, + { + "epoch": 31.05263157894737, + "grad_norm": 0.04624922573566437, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20650 + }, + { + "epoch": 31.06766917293233, + "grad_norm": 0.03547649830579758, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 20660 + }, + { + "epoch": 31.082706766917294, + "grad_norm": 0.03274167701601982, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20670 + }, + { + "epoch": 31.097744360902254, + "grad_norm": 0.03359862044453621, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 20680 + }, + { + "epoch": 31.112781954887218, + "grad_norm": 0.0320468433201313, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 20690 + }, + { + "epoch": 31.127819548872182, + "grad_norm": 0.03717431798577309, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20700 + }, + { + "epoch": 31.142857142857142, + "grad_norm": 0.03071301244199276, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20710 + }, + { + "epoch": 31.157894736842106, + "grad_norm": 0.03121928870677948, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 20720 + }, + { + "epoch": 31.172932330827066, + "grad_norm": 0.03071415238082409, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20730 + }, + { + "epoch": 31.18796992481203, + "grad_norm": 0.035135671496391296, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20740 + }, + { + "epoch": 31.203007518796994, + "grad_norm": 0.022921917960047722, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 20750 + }, + { + "epoch": 31.218045112781954, + "grad_norm": 0.023472385480999947, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20760 + }, + { + "epoch": 31.23308270676692, + "grad_norm": 0.03770725056529045, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20770 + }, + { + "epoch": 31.24812030075188, + "grad_norm": 0.03226419910788536, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20780 + }, + { + "epoch": 31.263157894736842, + "grad_norm": 0.037617918103933334, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 20790 + }, + { + "epoch": 31.278195488721803, + "grad_norm": 0.03209061920642853, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20800 + }, + { + "epoch": 31.293233082706767, + "grad_norm": 0.027157209813594818, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20810 + }, + { + "epoch": 31.30827067669173, + "grad_norm": 0.03473038598895073, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20820 + }, + { + "epoch": 31.32330827067669, + "grad_norm": 0.031824663281440735, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20830 + }, + { + "epoch": 31.338345864661655, + "grad_norm": 0.024446578696370125, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20840 + }, + { + "epoch": 31.353383458646615, + "grad_norm": 0.03053315170109272, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20850 + }, + { + "epoch": 31.36842105263158, + "grad_norm": 0.04006139934062958, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20860 + }, + { + "epoch": 31.383458646616543, + "grad_norm": 0.024767432361841202, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20870 + }, + { + "epoch": 31.398496240601503, + "grad_norm": 0.02873155288398266, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 20880 + }, + { + "epoch": 31.413533834586467, + "grad_norm": 0.027750737965106964, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20890 + }, + { + "epoch": 31.428571428571427, + "grad_norm": 0.04264703392982483, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20900 + }, + { + "epoch": 31.44360902255639, + "grad_norm": 0.04348491132259369, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20910 + }, + { + "epoch": 31.458646616541355, + "grad_norm": 0.031728748232126236, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20920 + }, + { + "epoch": 31.473684210526315, + "grad_norm": 0.028941979631781578, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 20930 + }, + { + "epoch": 31.48872180451128, + "grad_norm": 0.03833877667784691, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20940 + }, + { + "epoch": 31.50375939849624, + "grad_norm": 0.023916438221931458, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20950 + }, + { + "epoch": 31.518796992481203, + "grad_norm": 0.039876557886600494, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 20960 + }, + { + "epoch": 31.533834586466167, + "grad_norm": 0.0450846366584301, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 20970 + }, + { + "epoch": 31.548872180451127, + "grad_norm": 0.03148174658417702, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 20980 + }, + { + "epoch": 31.56390977443609, + "grad_norm": 0.028166072443127632, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 20990 + }, + { + "epoch": 31.57894736842105, + "grad_norm": 0.05433887615799904, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21000 + }, + { + "epoch": 31.57894736842105, + "eval_cer": 0.019822805917531577, + "eval_loss": 0.08885720372200012, + "eval_runtime": 160.9839, + "eval_samples_per_second": 99.557, + "eval_steps_per_second": 0.783, + "eval_wer": 0.06958879136988104, + "step": 21000 + }, + { + "epoch": 31.593984962406015, + "grad_norm": 0.02522432617843151, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21010 + }, + { + "epoch": 31.60902255639098, + "grad_norm": 0.03508957102894783, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21020 + }, + { + "epoch": 31.62406015037594, + "grad_norm": 0.04104788973927498, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21030 + }, + { + "epoch": 31.639097744360903, + "grad_norm": 0.04045097902417183, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21040 + }, + { + "epoch": 31.654135338345863, + "grad_norm": 0.04489412158727646, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 21050 + }, + { + "epoch": 31.669172932330827, + "grad_norm": 0.04312744736671448, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21060 + }, + { + "epoch": 31.68421052631579, + "grad_norm": 0.04382782429456711, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 21070 + }, + { + "epoch": 31.69924812030075, + "grad_norm": 0.03456209972500801, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21080 + }, + { + "epoch": 31.714285714285715, + "grad_norm": 0.03430342301726341, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21090 + }, + { + "epoch": 31.729323308270676, + "grad_norm": 0.03255769982933998, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 21100 + }, + { + "epoch": 31.74436090225564, + "grad_norm": 0.03125816211104393, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21110 + }, + { + "epoch": 31.7593984962406, + "grad_norm": 0.03934122249484062, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 21120 + }, + { + "epoch": 31.774436090225564, + "grad_norm": 0.035840362310409546, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21130 + }, + { + "epoch": 31.789473684210527, + "grad_norm": 1.7192158699035645, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 21140 + }, + { + "epoch": 31.804511278195488, + "grad_norm": 0.04431113973259926, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 21150 + }, + { + "epoch": 31.81954887218045, + "grad_norm": 0.0678129568696022, + "learning_rate": 0.0005, + "loss": 0.0033, + "step": 21160 + }, + { + "epoch": 31.834586466165412, + "grad_norm": 0.04205244779586792, + "learning_rate": 0.0005, + "loss": 0.0038, + "step": 21170 + }, + { + "epoch": 31.849624060150376, + "grad_norm": 0.03495098650455475, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21180 + }, + { + "epoch": 31.86466165413534, + "grad_norm": 0.041340406984090805, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 21190 + }, + { + "epoch": 31.8796992481203, + "grad_norm": 0.03226783126592636, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 21200 + }, + { + "epoch": 31.894736842105264, + "grad_norm": 0.04496467486023903, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21210 + }, + { + "epoch": 31.909774436090224, + "grad_norm": 0.04127941653132439, + "learning_rate": 0.0005, + "loss": 0.0035, + "step": 21220 + }, + { + "epoch": 31.924812030075188, + "grad_norm": 0.04743132367730141, + "learning_rate": 0.0005, + "loss": 0.0034, + "step": 21230 + }, + { + "epoch": 31.93984962406015, + "grad_norm": 0.02961602248251438, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 21240 + }, + { + "epoch": 31.954887218045112, + "grad_norm": 0.03538179397583008, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 21250 + }, + { + "epoch": 31.969924812030076, + "grad_norm": 0.03758075088262558, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 21260 + }, + { + "epoch": 31.984962406015036, + "grad_norm": 0.0232989601790905, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 21270 + }, + { + "epoch": 32.0, + "grad_norm": 0.027478547766804695, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21280 + }, + { + "epoch": 32.015037593984964, + "grad_norm": 0.04151836037635803, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21290 + }, + { + "epoch": 32.03007518796993, + "grad_norm": 0.03126261755824089, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21300 + }, + { + "epoch": 32.045112781954884, + "grad_norm": 0.02348332293331623, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21310 + }, + { + "epoch": 32.06015037593985, + "grad_norm": 0.04083660617470741, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21320 + }, + { + "epoch": 32.07518796992481, + "grad_norm": 0.03527889400720596, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21330 + }, + { + "epoch": 32.090225563909776, + "grad_norm": 0.037207480520009995, + "learning_rate": 0.0005, + "loss": 0.003, + "step": 21340 + }, + { + "epoch": 32.10526315789474, + "grad_norm": 0.03170577809214592, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21350 + }, + { + "epoch": 32.1203007518797, + "grad_norm": 0.03909881412982941, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21360 + }, + { + "epoch": 32.13533834586466, + "grad_norm": 0.04553251340985298, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21370 + }, + { + "epoch": 32.150375939849624, + "grad_norm": 0.030170215293765068, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 21380 + }, + { + "epoch": 32.16541353383459, + "grad_norm": 0.021695656701922417, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21390 + }, + { + "epoch": 32.18045112781955, + "grad_norm": 0.03792033717036247, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21400 + }, + { + "epoch": 32.19548872180451, + "grad_norm": 0.03648845851421356, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21410 + }, + { + "epoch": 32.21052631578947, + "grad_norm": 0.045861147344112396, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 21420 + }, + { + "epoch": 32.225563909774436, + "grad_norm": 0.03417893871665001, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21430 + }, + { + "epoch": 32.2406015037594, + "grad_norm": 0.0272667296230793, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21440 + }, + { + "epoch": 32.255639097744364, + "grad_norm": 0.04674240201711655, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21450 + }, + { + "epoch": 32.27067669172932, + "grad_norm": 0.03263983130455017, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21460 + }, + { + "epoch": 32.285714285714285, + "grad_norm": 0.029313135892152786, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21470 + }, + { + "epoch": 32.30075187969925, + "grad_norm": 0.025477949529886246, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21480 + }, + { + "epoch": 32.31578947368421, + "grad_norm": 0.04999217018485069, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21490 + }, + { + "epoch": 32.330827067669176, + "grad_norm": 0.034942325204610825, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21500 + }, + { + "epoch": 32.34586466165413, + "grad_norm": 0.01989215798676014, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21510 + }, + { + "epoch": 32.3609022556391, + "grad_norm": 0.030988946557044983, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 21520 + }, + { + "epoch": 32.37593984962406, + "grad_norm": 0.01521662063896656, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 21530 + }, + { + "epoch": 32.390977443609025, + "grad_norm": 0.02928643673658371, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 21540 + }, + { + "epoch": 32.40601503759399, + "grad_norm": 0.03286590427160263, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21550 + }, + { + "epoch": 32.421052631578945, + "grad_norm": 0.05697145313024521, + "learning_rate": 0.0005, + "loss": 0.0032, + "step": 21560 + }, + { + "epoch": 32.43609022556391, + "grad_norm": 0.01639280468225479, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21570 + }, + { + "epoch": 32.45112781954887, + "grad_norm": 0.03875429928302765, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 21580 + }, + { + "epoch": 32.46616541353384, + "grad_norm": 0.040002062916755676, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 21590 + }, + { + "epoch": 32.4812030075188, + "grad_norm": 0.029198426753282547, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21600 + }, + { + "epoch": 32.49624060150376, + "grad_norm": 0.026585113257169724, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 21610 + }, + { + "epoch": 32.51127819548872, + "grad_norm": 0.026823850348591805, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21620 + }, + { + "epoch": 32.526315789473685, + "grad_norm": 0.03076923079788685, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21630 + }, + { + "epoch": 32.54135338345865, + "grad_norm": 0.025849780067801476, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 21640 + }, + { + "epoch": 32.556390977443606, + "grad_norm": 0.052123699337244034, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21650 + }, + { + "epoch": 32.57142857142857, + "grad_norm": 0.05512465164065361, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21660 + }, + { + "epoch": 32.58646616541353, + "grad_norm": 0.040366966277360916, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21670 + }, + { + "epoch": 32.6015037593985, + "grad_norm": 0.04850524291396141, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21680 + }, + { + "epoch": 32.61654135338346, + "grad_norm": 0.03173189237713814, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21690 + }, + { + "epoch": 32.63157894736842, + "grad_norm": 0.03444327041506767, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21700 + }, + { + "epoch": 32.64661654135338, + "grad_norm": 0.021614478901028633, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21710 + }, + { + "epoch": 32.661654135338345, + "grad_norm": 0.03152332454919815, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21720 + }, + { + "epoch": 32.67669172932331, + "grad_norm": 0.04235197231173515, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 21730 + }, + { + "epoch": 32.69172932330827, + "grad_norm": 0.03243883699178696, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21740 + }, + { + "epoch": 32.70676691729323, + "grad_norm": 0.033237237483263016, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21750 + }, + { + "epoch": 32.721804511278194, + "grad_norm": 0.036359455436468124, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21760 + }, + { + "epoch": 32.73684210526316, + "grad_norm": 0.025052307173609734, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21770 + }, + { + "epoch": 32.75187969924812, + "grad_norm": 0.03170343488454819, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21780 + }, + { + "epoch": 32.766917293233085, + "grad_norm": 0.028873203322291374, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21790 + }, + { + "epoch": 32.78195488721804, + "grad_norm": 0.03198118507862091, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21800 + }, + { + "epoch": 32.796992481203006, + "grad_norm": 0.04021941125392914, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21810 + }, + { + "epoch": 32.81203007518797, + "grad_norm": 0.04307381063699722, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 21820 + }, + { + "epoch": 32.82706766917293, + "grad_norm": 0.04938417300581932, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21830 + }, + { + "epoch": 32.8421052631579, + "grad_norm": 0.049032412469387054, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21840 + }, + { + "epoch": 32.857142857142854, + "grad_norm": 0.029056649655103683, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21850 + }, + { + "epoch": 32.87218045112782, + "grad_norm": 0.047337066382169724, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21860 + }, + { + "epoch": 32.88721804511278, + "grad_norm": 0.02141435444355011, + "learning_rate": 0.0005, + "loss": 0.0028, + "step": 21870 + }, + { + "epoch": 32.902255639097746, + "grad_norm": 0.02494044229388237, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21880 + }, + { + "epoch": 32.91729323308271, + "grad_norm": 0.02970977872610092, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21890 + }, + { + "epoch": 32.932330827067666, + "grad_norm": 0.04080621153116226, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21900 + }, + { + "epoch": 32.94736842105263, + "grad_norm": 0.03223215043544769, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 21910 + }, + { + "epoch": 32.962406015037594, + "grad_norm": 0.036143708974123, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21920 + }, + { + "epoch": 32.97744360902256, + "grad_norm": 0.035308100283145905, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 21930 + }, + { + "epoch": 32.99248120300752, + "grad_norm": 0.03567254915833473, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 21940 + }, + { + "epoch": 33.00751879699248, + "grad_norm": 0.02538500353693962, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 21950 + }, + { + "epoch": 33.02255639097744, + "grad_norm": 0.024782000109553337, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 21960 + }, + { + "epoch": 33.037593984962406, + "grad_norm": 0.0399056002497673, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 21970 + }, + { + "epoch": 33.05263157894737, + "grad_norm": 0.02398042567074299, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 21980 + }, + { + "epoch": 33.067669172932334, + "grad_norm": 0.023772910237312317, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 21990 + }, + { + "epoch": 33.08270676691729, + "grad_norm": 0.02893906831741333, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22000 + }, + { + "epoch": 33.08270676691729, + "eval_cer": 0.019254283301076248, + "eval_loss": 0.09050923585891724, + "eval_runtime": 162.4766, + "eval_samples_per_second": 98.642, + "eval_steps_per_second": 0.775, + "eval_wer": 0.06860681463662162, + "step": 22000 + }, + { + "epoch": 33.097744360902254, + "grad_norm": 0.03247937932610512, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22010 + }, + { + "epoch": 33.11278195488722, + "grad_norm": 0.016984833404421806, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22020 + }, + { + "epoch": 33.12781954887218, + "grad_norm": 0.02986994758248329, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 22030 + }, + { + "epoch": 33.142857142857146, + "grad_norm": 0.03177684172987938, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22040 + }, + { + "epoch": 33.1578947368421, + "grad_norm": 0.02704494073987007, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22050 + }, + { + "epoch": 33.17293233082707, + "grad_norm": 0.03909464552998543, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22060 + }, + { + "epoch": 33.18796992481203, + "grad_norm": 0.02816700004041195, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 22070 + }, + { + "epoch": 33.203007518796994, + "grad_norm": 0.03229742869734764, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22080 + }, + { + "epoch": 33.21804511278196, + "grad_norm": 0.0349409393966198, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22090 + }, + { + "epoch": 33.233082706766915, + "grad_norm": 0.031935837119817734, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22100 + }, + { + "epoch": 33.24812030075188, + "grad_norm": 0.05277466028928757, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22110 + }, + { + "epoch": 33.26315789473684, + "grad_norm": 0.037275027483701706, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22120 + }, + { + "epoch": 33.278195488721806, + "grad_norm": 0.02872217632830143, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22130 + }, + { + "epoch": 33.29323308270677, + "grad_norm": 0.03234171122312546, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22140 + }, + { + "epoch": 33.30827067669173, + "grad_norm": 0.027111805975437164, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22150 + }, + { + "epoch": 33.32330827067669, + "grad_norm": 0.029638931155204773, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 22160 + }, + { + "epoch": 33.338345864661655, + "grad_norm": 0.035005196928977966, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22170 + }, + { + "epoch": 33.35338345864662, + "grad_norm": 0.02897673286497593, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22180 + }, + { + "epoch": 33.36842105263158, + "grad_norm": 0.027993859723210335, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 22190 + }, + { + "epoch": 33.38345864661654, + "grad_norm": 0.035929206758737564, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22200 + }, + { + "epoch": 33.3984962406015, + "grad_norm": 0.03395345062017441, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22210 + }, + { + "epoch": 33.41353383458647, + "grad_norm": 0.018052542582154274, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22220 + }, + { + "epoch": 33.42857142857143, + "grad_norm": 0.029163192957639694, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22230 + }, + { + "epoch": 33.443609022556394, + "grad_norm": 0.028594741597771645, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22240 + }, + { + "epoch": 33.45864661654135, + "grad_norm": 0.02453514188528061, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22250 + }, + { + "epoch": 33.473684210526315, + "grad_norm": 0.03020423837006092, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22260 + }, + { + "epoch": 33.48872180451128, + "grad_norm": 0.02218945138156414, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22270 + }, + { + "epoch": 33.50375939849624, + "grad_norm": 0.024073993787169456, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22280 + }, + { + "epoch": 33.5187969924812, + "grad_norm": 0.027833495289087296, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22290 + }, + { + "epoch": 33.53383458646616, + "grad_norm": 0.04346552491188049, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22300 + }, + { + "epoch": 33.54887218045113, + "grad_norm": 0.0446699894964695, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22310 + }, + { + "epoch": 33.56390977443609, + "grad_norm": 0.022739259526133537, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22320 + }, + { + "epoch": 33.578947368421055, + "grad_norm": 0.028683165088295937, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 22330 + }, + { + "epoch": 33.59398496240601, + "grad_norm": 0.026096219196915627, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22340 + }, + { + "epoch": 33.609022556390975, + "grad_norm": 0.028459154069423676, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22350 + }, + { + "epoch": 33.62406015037594, + "grad_norm": 0.038563426584005356, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22360 + }, + { + "epoch": 33.6390977443609, + "grad_norm": 0.022987637668848038, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22370 + }, + { + "epoch": 33.65413533834587, + "grad_norm": 0.029239581897854805, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22380 + }, + { + "epoch": 33.669172932330824, + "grad_norm": 0.02502906136214733, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22390 + }, + { + "epoch": 33.68421052631579, + "grad_norm": 0.03272206336259842, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22400 + }, + { + "epoch": 33.69924812030075, + "grad_norm": 0.026442624628543854, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22410 + }, + { + "epoch": 33.714285714285715, + "grad_norm": 0.033022891730070114, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22420 + }, + { + "epoch": 33.72932330827068, + "grad_norm": 0.036274682730436325, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22430 + }, + { + "epoch": 33.744360902255636, + "grad_norm": 0.022444242611527443, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22440 + }, + { + "epoch": 33.7593984962406, + "grad_norm": 0.03272925317287445, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 22450 + }, + { + "epoch": 33.774436090225564, + "grad_norm": 0.03297911584377289, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 22460 + }, + { + "epoch": 33.78947368421053, + "grad_norm": 0.03567880764603615, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22470 + }, + { + "epoch": 33.80451127819549, + "grad_norm": 0.030743472278118134, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22480 + }, + { + "epoch": 33.81954887218045, + "grad_norm": 0.0303607527166605, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22490 + }, + { + "epoch": 33.83458646616541, + "grad_norm": 0.061927102506160736, + "learning_rate": 0.0005, + "loss": 0.0029, + "step": 22500 + }, + { + "epoch": 33.849624060150376, + "grad_norm": 0.024137957021594048, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22510 + }, + { + "epoch": 33.86466165413534, + "grad_norm": 0.02823724038898945, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22520 + }, + { + "epoch": 33.8796992481203, + "grad_norm": 0.04236317053437233, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22530 + }, + { + "epoch": 33.89473684210526, + "grad_norm": 0.032932817935943604, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22540 + }, + { + "epoch": 33.909774436090224, + "grad_norm": 0.03780663385987282, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 22550 + }, + { + "epoch": 33.92481203007519, + "grad_norm": 0.04493261128664017, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 22560 + }, + { + "epoch": 33.93984962406015, + "grad_norm": 0.03775694593787193, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22570 + }, + { + "epoch": 33.954887218045116, + "grad_norm": 0.026718024164438248, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22580 + }, + { + "epoch": 33.96992481203007, + "grad_norm": 0.039048079401254654, + "learning_rate": 0.0005, + "loss": 0.0031, + "step": 22590 + }, + { + "epoch": 33.984962406015036, + "grad_norm": 0.024839913472533226, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22600 + }, + { + "epoch": 34.0, + "grad_norm": 0.03368949517607689, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22610 + }, + { + "epoch": 34.015037593984964, + "grad_norm": 0.02313544973731041, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22620 + }, + { + "epoch": 34.03007518796993, + "grad_norm": 0.033821966499090195, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 22630 + }, + { + "epoch": 34.045112781954884, + "grad_norm": 0.032729119062423706, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 22640 + }, + { + "epoch": 34.06015037593985, + "grad_norm": 0.04073363542556763, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22650 + }, + { + "epoch": 34.07518796992481, + "grad_norm": 0.03454175218939781, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22660 + }, + { + "epoch": 34.090225563909776, + "grad_norm": 0.025971870869398117, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22670 + }, + { + "epoch": 34.10526315789474, + "grad_norm": 0.02422664873301983, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22680 + }, + { + "epoch": 34.1203007518797, + "grad_norm": 0.03668517619371414, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22690 + }, + { + "epoch": 34.13533834586466, + "grad_norm": 0.02679029107093811, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22700 + }, + { + "epoch": 34.150375939849624, + "grad_norm": 0.03591066598892212, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22710 + }, + { + "epoch": 34.16541353383459, + "grad_norm": 0.0346236526966095, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22720 + }, + { + "epoch": 34.18045112781955, + "grad_norm": 0.02896529622375965, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22730 + }, + { + "epoch": 34.19548872180451, + "grad_norm": 0.025756575167179108, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22740 + }, + { + "epoch": 34.21052631578947, + "grad_norm": 0.02884383499622345, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 22750 + }, + { + "epoch": 34.225563909774436, + "grad_norm": 0.028614552691578865, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22760 + }, + { + "epoch": 34.2406015037594, + "grad_norm": 0.035526640713214874, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22770 + }, + { + "epoch": 34.255639097744364, + "grad_norm": 0.026349162682890892, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22780 + }, + { + "epoch": 34.27067669172932, + "grad_norm": 0.04058104008436203, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22790 + }, + { + "epoch": 34.285714285714285, + "grad_norm": 0.033807799220085144, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22800 + }, + { + "epoch": 34.30075187969925, + "grad_norm": 0.03374028950929642, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22810 + }, + { + "epoch": 34.31578947368421, + "grad_norm": 0.02964676544070244, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22820 + }, + { + "epoch": 34.330827067669176, + "grad_norm": 0.0487256795167923, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22830 + }, + { + "epoch": 34.34586466165413, + "grad_norm": 0.047146931290626526, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 22840 + }, + { + "epoch": 34.3609022556391, + "grad_norm": 0.02859164960682392, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22850 + }, + { + "epoch": 34.37593984962406, + "grad_norm": 0.027669129893183708, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 22860 + }, + { + "epoch": 34.390977443609025, + "grad_norm": 0.035878121852874756, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22870 + }, + { + "epoch": 34.40601503759399, + "grad_norm": 0.032285116612911224, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 22880 + }, + { + "epoch": 34.421052631578945, + "grad_norm": 0.046112217009067535, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 22890 + }, + { + "epoch": 34.43609022556391, + "grad_norm": 0.03079863451421261, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 22900 + }, + { + "epoch": 34.45112781954887, + "grad_norm": 0.03159705922007561, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22910 + }, + { + "epoch": 34.46616541353384, + "grad_norm": 0.041945986449718475, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 22920 + }, + { + "epoch": 34.4812030075188, + "grad_norm": 0.017488420009613037, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 22930 + }, + { + "epoch": 34.49624060150376, + "grad_norm": 0.03179776296019554, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22940 + }, + { + "epoch": 34.51127819548872, + "grad_norm": 0.023064451292157173, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22950 + }, + { + "epoch": 34.526315789473685, + "grad_norm": 0.03865617886185646, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 22960 + }, + { + "epoch": 34.54135338345865, + "grad_norm": 0.035459961742162704, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 22970 + }, + { + "epoch": 34.556390977443606, + "grad_norm": 0.038326773792505264, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 22980 + }, + { + "epoch": 34.57142857142857, + "grad_norm": 0.02682003565132618, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 22990 + }, + { + "epoch": 34.58646616541353, + "grad_norm": 0.021616095677018166, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23000 + }, + { + "epoch": 34.58646616541353, + "eval_cer": 0.019021705867071796, + "eval_loss": 0.09297627210617065, + "eval_runtime": 158.2894, + "eval_samples_per_second": 101.251, + "eval_steps_per_second": 0.796, + "eval_wer": 0.06883703884681162, + "step": 23000 + }, + { + "epoch": 34.6015037593985, + "grad_norm": 0.03111901879310608, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 23010 + }, + { + "epoch": 34.61654135338346, + "grad_norm": 0.02672715298831463, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23020 + }, + { + "epoch": 34.63157894736842, + "grad_norm": 0.028712518513202667, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23030 + }, + { + "epoch": 34.64661654135338, + "grad_norm": 0.03885149955749512, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23040 + }, + { + "epoch": 34.661654135338345, + "grad_norm": 0.029531456530094147, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23050 + }, + { + "epoch": 34.67669172932331, + "grad_norm": 0.027363164350390434, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23060 + }, + { + "epoch": 34.69172932330827, + "grad_norm": 0.023223329335451126, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23070 + }, + { + "epoch": 34.70676691729323, + "grad_norm": 0.036881223320961, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 23080 + }, + { + "epoch": 34.721804511278194, + "grad_norm": 0.023478113114833832, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23090 + }, + { + "epoch": 34.73684210526316, + "grad_norm": 0.02833597920835018, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 23100 + }, + { + "epoch": 34.75187969924812, + "grad_norm": 0.02242576889693737, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23110 + }, + { + "epoch": 34.766917293233085, + "grad_norm": 0.032836344093084335, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 23120 + }, + { + "epoch": 34.78195488721804, + "grad_norm": 0.035401564091444016, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 23130 + }, + { + "epoch": 34.796992481203006, + "grad_norm": 0.01934722065925598, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23140 + }, + { + "epoch": 34.81203007518797, + "grad_norm": 0.031650736927986145, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23150 + }, + { + "epoch": 34.82706766917293, + "grad_norm": 0.030918927863240242, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23160 + }, + { + "epoch": 34.8421052631579, + "grad_norm": 0.02688918262720108, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23170 + }, + { + "epoch": 34.857142857142854, + "grad_norm": 0.024265650659799576, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23180 + }, + { + "epoch": 34.87218045112782, + "grad_norm": 0.02701210044324398, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23190 + }, + { + "epoch": 34.88721804511278, + "grad_norm": 0.0323820486664772, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23200 + }, + { + "epoch": 34.902255639097746, + "grad_norm": 0.03181002661585808, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 23210 + }, + { + "epoch": 34.91729323308271, + "grad_norm": 0.041740454733371735, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 23220 + }, + { + "epoch": 34.932330827067666, + "grad_norm": 0.05627287179231644, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23230 + }, + { + "epoch": 34.94736842105263, + "grad_norm": 0.03051813691854477, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23240 + }, + { + "epoch": 34.962406015037594, + "grad_norm": 0.037858448922634125, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23250 + }, + { + "epoch": 34.97744360902256, + "grad_norm": 0.03268805518746376, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 23260 + }, + { + "epoch": 34.99248120300752, + "grad_norm": 0.0328117236495018, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 23270 + }, + { + "epoch": 35.00751879699248, + "grad_norm": 0.028408410027623177, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23280 + }, + { + "epoch": 35.02255639097744, + "grad_norm": 0.012817631475627422, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23290 + }, + { + "epoch": 35.037593984962406, + "grad_norm": 0.02666749618947506, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 23300 + }, + { + "epoch": 35.05263157894737, + "grad_norm": 0.04354560747742653, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23310 + }, + { + "epoch": 35.067669172932334, + "grad_norm": 0.06089859455823898, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 23320 + }, + { + "epoch": 35.08270676691729, + "grad_norm": 0.020502714440226555, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23330 + }, + { + "epoch": 35.097744360902254, + "grad_norm": 0.020051153376698494, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23340 + }, + { + "epoch": 35.11278195488722, + "grad_norm": 0.03633871674537659, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 23350 + }, + { + "epoch": 35.12781954887218, + "grad_norm": 0.0377647802233696, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 23360 + }, + { + "epoch": 35.142857142857146, + "grad_norm": 0.027437595650553703, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23370 + }, + { + "epoch": 35.1578947368421, + "grad_norm": 0.04558334872126579, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23380 + }, + { + "epoch": 35.17293233082707, + "grad_norm": 0.024032684043049812, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 23390 + }, + { + "epoch": 35.18796992481203, + "grad_norm": 0.02398781105875969, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23400 + }, + { + "epoch": 35.203007518796994, + "grad_norm": 0.019802546128630638, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23410 + }, + { + "epoch": 35.21804511278196, + "grad_norm": 0.03111514449119568, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23420 + }, + { + "epoch": 35.233082706766915, + "grad_norm": 0.0346931591629982, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23430 + }, + { + "epoch": 35.24812030075188, + "grad_norm": 0.022063296288251877, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23440 + }, + { + "epoch": 35.26315789473684, + "grad_norm": 0.02678515948355198, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23450 + }, + { + "epoch": 35.278195488721806, + "grad_norm": 0.02822466380894184, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23460 + }, + { + "epoch": 35.29323308270677, + "grad_norm": 0.023558564484119415, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23470 + }, + { + "epoch": 35.30827067669173, + "grad_norm": 0.04078399017453194, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23480 + }, + { + "epoch": 35.32330827067669, + "grad_norm": 0.030844811350107193, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23490 + }, + { + "epoch": 35.338345864661655, + "grad_norm": 0.026350408792495728, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23500 + }, + { + "epoch": 35.35338345864662, + "grad_norm": 0.030351990833878517, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23510 + }, + { + "epoch": 35.36842105263158, + "grad_norm": 0.027438346296548843, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23520 + }, + { + "epoch": 35.38345864661654, + "grad_norm": 0.03493416681885719, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23530 + }, + { + "epoch": 35.3984962406015, + "grad_norm": 0.026608053594827652, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23540 + }, + { + "epoch": 35.41353383458647, + "grad_norm": 0.02935125306248665, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 23550 + }, + { + "epoch": 35.42857142857143, + "grad_norm": 0.021243039518594742, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23560 + }, + { + "epoch": 35.443609022556394, + "grad_norm": 0.03764105588197708, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23570 + }, + { + "epoch": 35.45864661654135, + "grad_norm": 0.02640562690794468, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 23580 + }, + { + "epoch": 35.473684210526315, + "grad_norm": 0.029898643493652344, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23590 + }, + { + "epoch": 35.48872180451128, + "grad_norm": 0.02146708406507969, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23600 + }, + { + "epoch": 35.50375939849624, + "grad_norm": 0.027736341580748558, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 23610 + }, + { + "epoch": 35.5187969924812, + "grad_norm": 0.024983234703540802, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23620 + }, + { + "epoch": 35.53383458646616, + "grad_norm": 0.015204563736915588, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23630 + }, + { + "epoch": 35.54887218045113, + "grad_norm": 0.02606259658932686, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23640 + }, + { + "epoch": 35.56390977443609, + "grad_norm": 0.029908524826169014, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23650 + }, + { + "epoch": 35.578947368421055, + "grad_norm": 0.0301276333630085, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23660 + }, + { + "epoch": 35.59398496240601, + "grad_norm": 0.030795779079198837, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23670 + }, + { + "epoch": 35.609022556390975, + "grad_norm": 0.0317670963704586, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 23680 + }, + { + "epoch": 35.62406015037594, + "grad_norm": 0.022256160154938698, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23690 + }, + { + "epoch": 35.6390977443609, + "grad_norm": 0.04423995688557625, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23700 + }, + { + "epoch": 35.65413533834587, + "grad_norm": 0.038367755711078644, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23710 + }, + { + "epoch": 35.669172932330824, + "grad_norm": 0.02632117085158825, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23720 + }, + { + "epoch": 35.68421052631579, + "grad_norm": 0.027812309563159943, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23730 + }, + { + "epoch": 35.69924812030075, + "grad_norm": 0.01941763609647751, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23740 + }, + { + "epoch": 35.714285714285715, + "grad_norm": 0.04020305350422859, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23750 + }, + { + "epoch": 35.72932330827068, + "grad_norm": 0.035225238651037216, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23760 + }, + { + "epoch": 35.744360902255636, + "grad_norm": 0.03685884550213814, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23770 + }, + { + "epoch": 35.7593984962406, + "grad_norm": 0.019977448508143425, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23780 + }, + { + "epoch": 35.774436090225564, + "grad_norm": 0.023180020973086357, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23790 + }, + { + "epoch": 35.78947368421053, + "grad_norm": 0.027110040187835693, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23800 + }, + { + "epoch": 35.80451127819549, + "grad_norm": 0.03266303986310959, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23810 + }, + { + "epoch": 35.81954887218045, + "grad_norm": 0.022732997313141823, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23820 + }, + { + "epoch": 35.83458646616541, + "grad_norm": 0.018934201449155807, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 23830 + }, + { + "epoch": 35.849624060150376, + "grad_norm": 0.019439229741692543, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23840 + }, + { + "epoch": 35.86466165413534, + "grad_norm": 0.015423495322465897, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23850 + }, + { + "epoch": 35.8796992481203, + "grad_norm": 0.03718620538711548, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 23860 + }, + { + "epoch": 35.89473684210526, + "grad_norm": 0.03314783424139023, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23870 + }, + { + "epoch": 35.909774436090224, + "grad_norm": 0.018321670591831207, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23880 + }, + { + "epoch": 35.92481203007519, + "grad_norm": 0.027145950123667717, + "learning_rate": 0.0005, + "loss": 0.0027, + "step": 23890 + }, + { + "epoch": 35.93984962406015, + "grad_norm": 0.021608775481581688, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23900 + }, + { + "epoch": 35.954887218045116, + "grad_norm": 0.02125662751495838, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 23910 + }, + { + "epoch": 35.96992481203007, + "grad_norm": 0.028965091332793236, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23920 + }, + { + "epoch": 35.984962406015036, + "grad_norm": 0.02480863407254219, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 23930 + }, + { + "epoch": 36.0, + "grad_norm": 0.039592258632183075, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 23940 + }, + { + "epoch": 36.015037593984964, + "grad_norm": 0.026441309601068497, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23950 + }, + { + "epoch": 36.03007518796993, + "grad_norm": 0.016463376581668854, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 23960 + }, + { + "epoch": 36.045112781954884, + "grad_norm": 0.03426481410861015, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 23970 + }, + { + "epoch": 36.06015037593985, + "grad_norm": 0.03713420405983925, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 23980 + }, + { + "epoch": 36.07518796992481, + "grad_norm": 0.022133640944957733, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 23990 + }, + { + "epoch": 36.090225563909776, + "grad_norm": 0.03143461421132088, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24000 + }, + { + "epoch": 36.090225563909776, + "eval_cer": 0.019505902162426095, + "eval_loss": 0.09578326344490051, + "eval_runtime": 161.6731, + "eval_samples_per_second": 99.132, + "eval_steps_per_second": 0.779, + "eval_wer": 0.0690484692439249, + "step": 24000 + }, + { + "epoch": 36.10526315789474, + "grad_norm": 0.028687598183751106, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24010 + }, + { + "epoch": 36.1203007518797, + "grad_norm": 0.0321597158908844, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24020 + }, + { + "epoch": 36.13533834586466, + "grad_norm": 0.0420401394367218, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24030 + }, + { + "epoch": 36.150375939849624, + "grad_norm": 0.02434290014207363, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24040 + }, + { + "epoch": 36.16541353383459, + "grad_norm": 0.03881194442510605, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 24050 + }, + { + "epoch": 36.18045112781955, + "grad_norm": 0.030382605269551277, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24060 + }, + { + "epoch": 36.19548872180451, + "grad_norm": 0.05196426808834076, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24070 + }, + { + "epoch": 36.21052631578947, + "grad_norm": 0.03176790475845337, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24080 + }, + { + "epoch": 36.225563909774436, + "grad_norm": 0.029753653332591057, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24090 + }, + { + "epoch": 36.2406015037594, + "grad_norm": 0.024460267275571823, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24100 + }, + { + "epoch": 36.255639097744364, + "grad_norm": 0.02604435198009014, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24110 + }, + { + "epoch": 36.27067669172932, + "grad_norm": 0.024882299825549126, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24120 + }, + { + "epoch": 36.285714285714285, + "grad_norm": 0.01922144740819931, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 24130 + }, + { + "epoch": 36.30075187969925, + "grad_norm": 0.02054821513593197, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24140 + }, + { + "epoch": 36.31578947368421, + "grad_norm": 0.019630759954452515, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24150 + }, + { + "epoch": 36.330827067669176, + "grad_norm": 0.031896937638521194, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24160 + }, + { + "epoch": 36.34586466165413, + "grad_norm": 0.03765524923801422, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 24170 + }, + { + "epoch": 36.3609022556391, + "grad_norm": 0.03559780493378639, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24180 + }, + { + "epoch": 36.37593984962406, + "grad_norm": 0.03859793394804001, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24190 + }, + { + "epoch": 36.390977443609025, + "grad_norm": 0.030556129291653633, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24200 + }, + { + "epoch": 36.40601503759399, + "grad_norm": 0.03387856110930443, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24210 + }, + { + "epoch": 36.421052631578945, + "grad_norm": 0.04557717218995094, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24220 + }, + { + "epoch": 36.43609022556391, + "grad_norm": 0.029402272775769234, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24230 + }, + { + "epoch": 36.45112781954887, + "grad_norm": 0.01850428804755211, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24240 + }, + { + "epoch": 36.46616541353384, + "grad_norm": 0.033334147185087204, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24250 + }, + { + "epoch": 36.4812030075188, + "grad_norm": 0.033912863582372665, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24260 + }, + { + "epoch": 36.49624060150376, + "grad_norm": 0.031264156103134155, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24270 + }, + { + "epoch": 36.51127819548872, + "grad_norm": 0.03146151080727577, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24280 + }, + { + "epoch": 36.526315789473685, + "grad_norm": 0.02323835901916027, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24290 + }, + { + "epoch": 36.54135338345865, + "grad_norm": 0.021720755845308304, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24300 + }, + { + "epoch": 36.556390977443606, + "grad_norm": 0.04202906787395477, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24310 + }, + { + "epoch": 36.57142857142857, + "grad_norm": 0.019227536395192146, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24320 + }, + { + "epoch": 36.58646616541353, + "grad_norm": 0.02790912799537182, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24330 + }, + { + "epoch": 36.6015037593985, + "grad_norm": 0.035885684192180634, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24340 + }, + { + "epoch": 36.61654135338346, + "grad_norm": 0.044938258826732635, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24350 + }, + { + "epoch": 36.63157894736842, + "grad_norm": 0.02873925305902958, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24360 + }, + { + "epoch": 36.64661654135338, + "grad_norm": 0.028431184589862823, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24370 + }, + { + "epoch": 36.661654135338345, + "grad_norm": 0.034783437848091125, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 24380 + }, + { + "epoch": 36.67669172932331, + "grad_norm": 0.033178482204675674, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24390 + }, + { + "epoch": 36.69172932330827, + "grad_norm": 0.025607194751501083, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24400 + }, + { + "epoch": 36.70676691729323, + "grad_norm": 0.018107280135154724, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24410 + }, + { + "epoch": 36.721804511278194, + "grad_norm": 0.027150632813572884, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24420 + }, + { + "epoch": 36.73684210526316, + "grad_norm": 0.027938470244407654, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24430 + }, + { + "epoch": 36.75187969924812, + "grad_norm": 0.019404198974370956, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24440 + }, + { + "epoch": 36.766917293233085, + "grad_norm": 0.03735586628317833, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 24450 + }, + { + "epoch": 36.78195488721804, + "grad_norm": 0.016296228393912315, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24460 + }, + { + "epoch": 36.796992481203006, + "grad_norm": 0.016996650025248528, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 24470 + }, + { + "epoch": 36.81203007518797, + "grad_norm": 0.02575741708278656, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24480 + }, + { + "epoch": 36.82706766917293, + "grad_norm": 0.04499669745564461, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24490 + }, + { + "epoch": 36.8421052631579, + "grad_norm": 0.026319274678826332, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24500 + }, + { + "epoch": 36.857142857142854, + "grad_norm": 0.036399200558662415, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24510 + }, + { + "epoch": 36.87218045112782, + "grad_norm": 0.03955905884504318, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24520 + }, + { + "epoch": 36.88721804511278, + "grad_norm": 0.03508684039115906, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 24530 + }, + { + "epoch": 36.902255639097746, + "grad_norm": 0.03212455287575722, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24540 + }, + { + "epoch": 36.91729323308271, + "grad_norm": 0.026980960741639137, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24550 + }, + { + "epoch": 36.932330827067666, + "grad_norm": 0.040985945612192154, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24560 + }, + { + "epoch": 36.94736842105263, + "grad_norm": 0.025940587744116783, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24570 + }, + { + "epoch": 36.962406015037594, + "grad_norm": 0.04203914850950241, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24580 + }, + { + "epoch": 36.97744360902256, + "grad_norm": 0.02406606450676918, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24590 + }, + { + "epoch": 36.99248120300752, + "grad_norm": 0.03130471333861351, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24600 + }, + { + "epoch": 37.00751879699248, + "grad_norm": 0.01941523887217045, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24610 + }, + { + "epoch": 37.02255639097744, + "grad_norm": 0.05118865147233009, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24620 + }, + { + "epoch": 37.037593984962406, + "grad_norm": 0.041613783687353134, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24630 + }, + { + "epoch": 37.05263157894737, + "grad_norm": 0.04548795893788338, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24640 + }, + { + "epoch": 37.067669172932334, + "grad_norm": 0.023476071655750275, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24650 + }, + { + "epoch": 37.08270676691729, + "grad_norm": 0.031952228397130966, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24660 + }, + { + "epoch": 37.097744360902254, + "grad_norm": 0.024584120139479637, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24670 + }, + { + "epoch": 37.11278195488722, + "grad_norm": 0.024230752140283585, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24680 + }, + { + "epoch": 37.12781954887218, + "grad_norm": 0.04538040980696678, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 24690 + }, + { + "epoch": 37.142857142857146, + "grad_norm": 0.028469275683164597, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24700 + }, + { + "epoch": 37.1578947368421, + "grad_norm": 0.031096121296286583, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24710 + }, + { + "epoch": 37.17293233082707, + "grad_norm": 0.049959369003772736, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24720 + }, + { + "epoch": 37.18796992481203, + "grad_norm": 0.02693651244044304, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 24730 + }, + { + "epoch": 37.203007518796994, + "grad_norm": 0.03144918382167816, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 24740 + }, + { + "epoch": 37.21804511278196, + "grad_norm": 0.029649754986166954, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24750 + }, + { + "epoch": 37.233082706766915, + "grad_norm": 0.032893989235162735, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24760 + }, + { + "epoch": 37.24812030075188, + "grad_norm": 0.025503195822238922, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24770 + }, + { + "epoch": 37.26315789473684, + "grad_norm": 0.03800758346915245, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24780 + }, + { + "epoch": 37.278195488721806, + "grad_norm": 0.048395104706287384, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24790 + }, + { + "epoch": 37.29323308270677, + "grad_norm": 0.020913390442728996, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24800 + }, + { + "epoch": 37.30827067669173, + "grad_norm": 0.026472685858607292, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24810 + }, + { + "epoch": 37.32330827067669, + "grad_norm": 0.025544729083776474, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24820 + }, + { + "epoch": 37.338345864661655, + "grad_norm": 0.020984498783946037, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24830 + }, + { + "epoch": 37.35338345864662, + "grad_norm": 0.025402233004570007, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 24840 + }, + { + "epoch": 37.36842105263158, + "grad_norm": 0.04069389030337334, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24850 + }, + { + "epoch": 37.38345864661654, + "grad_norm": 0.02351643517613411, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24860 + }, + { + "epoch": 37.3984962406015, + "grad_norm": 0.024619558826088905, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24870 + }, + { + "epoch": 37.41353383458647, + "grad_norm": 0.027044691145420074, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24880 + }, + { + "epoch": 37.42857142857143, + "grad_norm": 0.027400491759181023, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 24890 + }, + { + "epoch": 37.443609022556394, + "grad_norm": 0.033194124698638916, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 24900 + }, + { + "epoch": 37.45864661654135, + "grad_norm": 0.02802683226764202, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 24910 + }, + { + "epoch": 37.473684210526315, + "grad_norm": 0.030347183346748352, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 24920 + }, + { + "epoch": 37.48872180451128, + "grad_norm": 0.038948871195316315, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 24930 + }, + { + "epoch": 37.50375939849624, + "grad_norm": 0.021162211894989014, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 24940 + }, + { + "epoch": 37.5187969924812, + "grad_norm": 0.044455770403146744, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24950 + }, + { + "epoch": 37.53383458646616, + "grad_norm": 0.026182519271969795, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 24960 + }, + { + "epoch": 37.54887218045113, + "grad_norm": 0.04024651646614075, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24970 + }, + { + "epoch": 37.56390977443609, + "grad_norm": 0.031114110723137856, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 24980 + }, + { + "epoch": 37.578947368421055, + "grad_norm": 0.027885517105460167, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 24990 + }, + { + "epoch": 37.59398496240601, + "grad_norm": 0.03926587849855423, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25000 + }, + { + "epoch": 37.59398496240601, + "eval_cer": 0.01902986647879125, + "eval_loss": 0.09147302061319351, + "eval_runtime": 158.8157, + "eval_samples_per_second": 100.916, + "eval_steps_per_second": 0.793, + "eval_wer": 0.06739931214644139, + "step": 25000 + }, + { + "epoch": 37.609022556390975, + "grad_norm": 0.02450677938759327, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25010 + }, + { + "epoch": 37.62406015037594, + "grad_norm": 0.02565346471965313, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25020 + }, + { + "epoch": 37.6390977443609, + "grad_norm": 0.018661195412278175, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25030 + }, + { + "epoch": 37.65413533834587, + "grad_norm": 0.018631530925631523, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25040 + }, + { + "epoch": 37.669172932330824, + "grad_norm": 0.023524627089500427, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25050 + }, + { + "epoch": 37.68421052631579, + "grad_norm": 0.033701151609420776, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25060 + }, + { + "epoch": 37.69924812030075, + "grad_norm": 0.024308161810040474, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25070 + }, + { + "epoch": 37.714285714285715, + "grad_norm": 0.02261587232351303, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25080 + }, + { + "epoch": 37.72932330827068, + "grad_norm": 0.050358157604932785, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 25090 + }, + { + "epoch": 37.744360902255636, + "grad_norm": 0.03883576765656471, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25100 + }, + { + "epoch": 37.7593984962406, + "grad_norm": 0.01848018541932106, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25110 + }, + { + "epoch": 37.774436090225564, + "grad_norm": 0.020785657688975334, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25120 + }, + { + "epoch": 37.78947368421053, + "grad_norm": 0.02417186088860035, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25130 + }, + { + "epoch": 37.80451127819549, + "grad_norm": 0.023672234266996384, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25140 + }, + { + "epoch": 37.81954887218045, + "grad_norm": 0.034534208476543427, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25150 + }, + { + "epoch": 37.83458646616541, + "grad_norm": 0.021914709359407425, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25160 + }, + { + "epoch": 37.849624060150376, + "grad_norm": 0.034332964569330215, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 25170 + }, + { + "epoch": 37.86466165413534, + "grad_norm": 0.041687965393066406, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25180 + }, + { + "epoch": 37.8796992481203, + "grad_norm": 0.033789440989494324, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 25190 + }, + { + "epoch": 37.89473684210526, + "grad_norm": 0.020379576832056046, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25200 + }, + { + "epoch": 37.909774436090224, + "grad_norm": 0.019564269110560417, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25210 + }, + { + "epoch": 37.92481203007519, + "grad_norm": 0.03583402931690216, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25220 + }, + { + "epoch": 37.93984962406015, + "grad_norm": 0.023064322769641876, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25230 + }, + { + "epoch": 37.954887218045116, + "grad_norm": 0.022204279899597168, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25240 + }, + { + "epoch": 37.96992481203007, + "grad_norm": 0.021162860095500946, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25250 + }, + { + "epoch": 37.984962406015036, + "grad_norm": 0.03853774443268776, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25260 + }, + { + "epoch": 38.0, + "grad_norm": 0.035942934453487396, + "learning_rate": 0.0005, + "loss": 0.0025, + "step": 25270 + }, + { + "epoch": 38.015037593984964, + "grad_norm": 0.01774127222597599, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25280 + }, + { + "epoch": 38.03007518796993, + "grad_norm": 0.030406462028622627, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25290 + }, + { + "epoch": 38.045112781954884, + "grad_norm": 0.021693676710128784, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25300 + }, + { + "epoch": 38.06015037593985, + "grad_norm": 0.019597910344600677, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 25310 + }, + { + "epoch": 38.07518796992481, + "grad_norm": 0.033640988171100616, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25320 + }, + { + "epoch": 38.090225563909776, + "grad_norm": 0.016898533329367638, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25330 + }, + { + "epoch": 38.10526315789474, + "grad_norm": 0.01724872551858425, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25340 + }, + { + "epoch": 38.1203007518797, + "grad_norm": 0.026744499802589417, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25350 + }, + { + "epoch": 38.13533834586466, + "grad_norm": 0.03026716224849224, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25360 + }, + { + "epoch": 38.150375939849624, + "grad_norm": 0.03448054939508438, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25370 + }, + { + "epoch": 38.16541353383459, + "grad_norm": 0.020261602476239204, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25380 + }, + { + "epoch": 38.18045112781955, + "grad_norm": 0.020617710426449776, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25390 + }, + { + "epoch": 38.19548872180451, + "grad_norm": 0.02845141477882862, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25400 + }, + { + "epoch": 38.21052631578947, + "grad_norm": 0.02863137423992157, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25410 + }, + { + "epoch": 38.225563909774436, + "grad_norm": 0.03827967122197151, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25420 + }, + { + "epoch": 38.2406015037594, + "grad_norm": 0.018661916255950928, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 25430 + }, + { + "epoch": 38.255639097744364, + "grad_norm": 0.022926293313503265, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 25440 + }, + { + "epoch": 38.27067669172932, + "grad_norm": 0.040309883654117584, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25450 + }, + { + "epoch": 38.285714285714285, + "grad_norm": 0.03215287625789642, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25460 + }, + { + "epoch": 38.30075187969925, + "grad_norm": 0.03568839281797409, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25470 + }, + { + "epoch": 38.31578947368421, + "grad_norm": 0.024662336334586143, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25480 + }, + { + "epoch": 38.330827067669176, + "grad_norm": 0.014395013451576233, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25490 + }, + { + "epoch": 38.34586466165413, + "grad_norm": 0.030174007639288902, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25500 + }, + { + "epoch": 38.3609022556391, + "grad_norm": 0.02336134761571884, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25510 + }, + { + "epoch": 38.37593984962406, + "grad_norm": 0.0293938796967268, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25520 + }, + { + "epoch": 38.390977443609025, + "grad_norm": 0.02456602454185486, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25530 + }, + { + "epoch": 38.40601503759399, + "grad_norm": 0.03239433467388153, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25540 + }, + { + "epoch": 38.421052631578945, + "grad_norm": 0.03092825412750244, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 25550 + }, + { + "epoch": 38.43609022556391, + "grad_norm": 0.0255194753408432, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25560 + }, + { + "epoch": 38.45112781954887, + "grad_norm": 0.04706099256873131, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25570 + }, + { + "epoch": 38.46616541353384, + "grad_norm": 0.026288719847798347, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 25580 + }, + { + "epoch": 38.4812030075188, + "grad_norm": 0.025913279503583908, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25590 + }, + { + "epoch": 38.49624060150376, + "grad_norm": 0.020595241338014603, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25600 + }, + { + "epoch": 38.51127819548872, + "grad_norm": 0.030422385782003403, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25610 + }, + { + "epoch": 38.526315789473685, + "grad_norm": 0.0279834046959877, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25620 + }, + { + "epoch": 38.54135338345865, + "grad_norm": 0.03373396396636963, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25630 + }, + { + "epoch": 38.556390977443606, + "grad_norm": 0.03706225007772446, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25640 + }, + { + "epoch": 38.57142857142857, + "grad_norm": 0.027588170021772385, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25650 + }, + { + "epoch": 38.58646616541353, + "grad_norm": 0.0308929942548275, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25660 + }, + { + "epoch": 38.6015037593985, + "grad_norm": 0.028206566348671913, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 25670 + }, + { + "epoch": 38.61654135338346, + "grad_norm": 0.0198623389005661, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25680 + }, + { + "epoch": 38.63157894736842, + "grad_norm": 0.022228408604860306, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25690 + }, + { + "epoch": 38.64661654135338, + "grad_norm": 0.033929865807294846, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25700 + }, + { + "epoch": 38.661654135338345, + "grad_norm": 0.027096863836050034, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25710 + }, + { + "epoch": 38.67669172932331, + "grad_norm": 0.02119867503643036, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25720 + }, + { + "epoch": 38.69172932330827, + "grad_norm": 0.04341566935181618, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25730 + }, + { + "epoch": 38.70676691729323, + "grad_norm": 0.013764388859272003, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25740 + }, + { + "epoch": 38.721804511278194, + "grad_norm": 0.02024315483868122, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25750 + }, + { + "epoch": 38.73684210526316, + "grad_norm": 0.02229958400130272, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25760 + }, + { + "epoch": 38.75187969924812, + "grad_norm": 0.027075855061411858, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25770 + }, + { + "epoch": 38.766917293233085, + "grad_norm": 0.02672194316983223, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25780 + }, + { + "epoch": 38.78195488721804, + "grad_norm": 0.04474640637636185, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25790 + }, + { + "epoch": 38.796992481203006, + "grad_norm": 0.031444206833839417, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 25800 + }, + { + "epoch": 38.81203007518797, + "grad_norm": 0.030260471627116203, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25810 + }, + { + "epoch": 38.82706766917293, + "grad_norm": 0.030382193624973297, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25820 + }, + { + "epoch": 38.8421052631579, + "grad_norm": 0.018475910648703575, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25830 + }, + { + "epoch": 38.857142857142854, + "grad_norm": 0.03264062851667404, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25840 + }, + { + "epoch": 38.87218045112782, + "grad_norm": 0.06608446687459946, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 25850 + }, + { + "epoch": 38.88721804511278, + "grad_norm": 0.042255550622940063, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 25860 + }, + { + "epoch": 38.902255639097746, + "grad_norm": 0.04312232881784439, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 25870 + }, + { + "epoch": 38.91729323308271, + "grad_norm": 0.02113906852900982, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25880 + }, + { + "epoch": 38.932330827067666, + "grad_norm": 0.03629100322723389, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25890 + }, + { + "epoch": 38.94736842105263, + "grad_norm": 0.027212055400013924, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25900 + }, + { + "epoch": 38.962406015037594, + "grad_norm": 0.035533055663108826, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 25910 + }, + { + "epoch": 38.97744360902256, + "grad_norm": 0.029517196118831635, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25920 + }, + { + "epoch": 38.99248120300752, + "grad_norm": 0.024982089176774025, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 25930 + }, + { + "epoch": 39.00751879699248, + "grad_norm": 0.03406618908047676, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 25940 + }, + { + "epoch": 39.02255639097744, + "grad_norm": 0.04072027653455734, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25950 + }, + { + "epoch": 39.037593984962406, + "grad_norm": 0.017215779051184654, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 25960 + }, + { + "epoch": 39.05263157894737, + "grad_norm": 0.025155793875455856, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 25970 + }, + { + "epoch": 39.067669172932334, + "grad_norm": 0.026057902723550797, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 25980 + }, + { + "epoch": 39.08270676691729, + "grad_norm": 0.03795222193002701, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 25990 + }, + { + "epoch": 39.097744360902254, + "grad_norm": 0.022597817704081535, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26000 + }, + { + "epoch": 39.097744360902254, + "eval_cer": 0.01845658350549957, + "eval_loss": 0.09023154526948929, + "eval_runtime": 160.3951, + "eval_samples_per_second": 99.922, + "eval_steps_per_second": 0.786, + "eval_wer": 0.06767182243605405, + "step": 26000 + }, + { + "epoch": 39.11278195488722, + "grad_norm": 0.027854323387145996, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26010 + }, + { + "epoch": 39.12781954887218, + "grad_norm": 0.028090883046388626, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26020 + }, + { + "epoch": 39.142857142857146, + "grad_norm": 0.045079391449689865, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26030 + }, + { + "epoch": 39.1578947368421, + "grad_norm": 0.04673139750957489, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26040 + }, + { + "epoch": 39.17293233082707, + "grad_norm": 0.06251508742570877, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26050 + }, + { + "epoch": 39.18796992481203, + "grad_norm": 0.019558705389499664, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26060 + }, + { + "epoch": 39.203007518796994, + "grad_norm": 0.039357829838991165, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26070 + }, + { + "epoch": 39.21804511278196, + "grad_norm": 0.03450632840394974, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26080 + }, + { + "epoch": 39.233082706766915, + "grad_norm": 0.028345460072159767, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26090 + }, + { + "epoch": 39.24812030075188, + "grad_norm": 0.04065714031457901, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26100 + }, + { + "epoch": 39.26315789473684, + "grad_norm": 0.03303253650665283, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26110 + }, + { + "epoch": 39.278195488721806, + "grad_norm": 0.025598177686333656, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26120 + }, + { + "epoch": 39.29323308270677, + "grad_norm": 0.030943188816308975, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26130 + }, + { + "epoch": 39.30827067669173, + "grad_norm": 0.028286339715123177, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 26140 + }, + { + "epoch": 39.32330827067669, + "grad_norm": 0.035792816430330276, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26150 + }, + { + "epoch": 39.338345864661655, + "grad_norm": 0.02355850487947464, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26160 + }, + { + "epoch": 39.35338345864662, + "grad_norm": 0.016475237905979156, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26170 + }, + { + "epoch": 39.36842105263158, + "grad_norm": 0.022354811429977417, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26180 + }, + { + "epoch": 39.38345864661654, + "grad_norm": 0.021933553740382195, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26190 + }, + { + "epoch": 39.3984962406015, + "grad_norm": 0.032341040670871735, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26200 + }, + { + "epoch": 39.41353383458647, + "grad_norm": 0.030540546402335167, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26210 + }, + { + "epoch": 39.42857142857143, + "grad_norm": 0.0803193673491478, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26220 + }, + { + "epoch": 39.443609022556394, + "grad_norm": 0.033865053206682205, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26230 + }, + { + "epoch": 39.45864661654135, + "grad_norm": 0.02584117464721203, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26240 + }, + { + "epoch": 39.473684210526315, + "grad_norm": 0.03289191424846649, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26250 + }, + { + "epoch": 39.48872180451128, + "grad_norm": 0.03070538304746151, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26260 + }, + { + "epoch": 39.50375939849624, + "grad_norm": 0.032999977469444275, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26270 + }, + { + "epoch": 39.5187969924812, + "grad_norm": 0.03349313884973526, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26280 + }, + { + "epoch": 39.53383458646616, + "grad_norm": 0.025797231122851372, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26290 + }, + { + "epoch": 39.54887218045113, + "grad_norm": 0.02313665859401226, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 26300 + }, + { + "epoch": 39.56390977443609, + "grad_norm": 0.0352800115942955, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26310 + }, + { + "epoch": 39.578947368421055, + "grad_norm": 0.03971441462635994, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26320 + }, + { + "epoch": 39.59398496240601, + "grad_norm": 0.026513898745179176, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26330 + }, + { + "epoch": 39.609022556390975, + "grad_norm": 0.0329996794462204, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26340 + }, + { + "epoch": 39.62406015037594, + "grad_norm": 0.026981886476278305, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26350 + }, + { + "epoch": 39.6390977443609, + "grad_norm": 0.017976095899939537, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26360 + }, + { + "epoch": 39.65413533834587, + "grad_norm": 0.02137616090476513, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26370 + }, + { + "epoch": 39.669172932330824, + "grad_norm": 0.030614536255598068, + "learning_rate": 0.0005, + "loss": 0.0026, + "step": 26380 + }, + { + "epoch": 39.68421052631579, + "grad_norm": 0.026190919801592827, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26390 + }, + { + "epoch": 39.69924812030075, + "grad_norm": 0.03509816527366638, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26400 + }, + { + "epoch": 39.714285714285715, + "grad_norm": 0.0193096324801445, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26410 + }, + { + "epoch": 39.72932330827068, + "grad_norm": 0.028398791328072548, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26420 + }, + { + "epoch": 39.744360902255636, + "grad_norm": 0.024091795086860657, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 26430 + }, + { + "epoch": 39.7593984962406, + "grad_norm": 0.025807760655879974, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26440 + }, + { + "epoch": 39.774436090225564, + "grad_norm": 0.026838194578886032, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26450 + }, + { + "epoch": 39.78947368421053, + "grad_norm": 0.025380630046129227, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26460 + }, + { + "epoch": 39.80451127819549, + "grad_norm": 0.024630431085824966, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 26470 + }, + { + "epoch": 39.81954887218045, + "grad_norm": 0.03176168352365494, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26480 + }, + { + "epoch": 39.83458646616541, + "grad_norm": 0.06626243889331818, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26490 + }, + { + "epoch": 39.849624060150376, + "grad_norm": 0.032029107213020325, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26500 + }, + { + "epoch": 39.86466165413534, + "grad_norm": 0.036308951675891876, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26510 + }, + { + "epoch": 39.8796992481203, + "grad_norm": 0.02596602775156498, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26520 + }, + { + "epoch": 39.89473684210526, + "grad_norm": 0.02088405191898346, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 26530 + }, + { + "epoch": 39.909774436090224, + "grad_norm": 0.024560507386922836, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26540 + }, + { + "epoch": 39.92481203007519, + "grad_norm": 0.03970738872885704, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26550 + }, + { + "epoch": 39.93984962406015, + "grad_norm": 0.044390566647052765, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26560 + }, + { + "epoch": 39.954887218045116, + "grad_norm": 0.024599477648735046, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26570 + }, + { + "epoch": 39.96992481203007, + "grad_norm": 0.0410950742661953, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 26580 + }, + { + "epoch": 39.984962406015036, + "grad_norm": 0.02761153317987919, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26590 + }, + { + "epoch": 40.0, + "grad_norm": 0.03531324863433838, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26600 + }, + { + "epoch": 40.015037593984964, + "grad_norm": 0.01470787450671196, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 26610 + }, + { + "epoch": 40.03007518796993, + "grad_norm": 0.021041158586740494, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 26620 + }, + { + "epoch": 40.045112781954884, + "grad_norm": 0.02675044536590576, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 26630 + }, + { + "epoch": 40.06015037593985, + "grad_norm": 0.02389891818165779, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 26640 + }, + { + "epoch": 40.07518796992481, + "grad_norm": 0.017203668132424355, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 26650 + }, + { + "epoch": 40.090225563909776, + "grad_norm": 0.03339096158742905, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 26660 + }, + { + "epoch": 40.10526315789474, + "grad_norm": 0.06738464534282684, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26670 + }, + { + "epoch": 40.1203007518797, + "grad_norm": 0.030031926929950714, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26680 + }, + { + "epoch": 40.13533834586466, + "grad_norm": 0.02415679395198822, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 26690 + }, + { + "epoch": 40.150375939849624, + "grad_norm": 0.02785368077456951, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 26700 + }, + { + "epoch": 40.16541353383459, + "grad_norm": 0.03518250957131386, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26710 + }, + { + "epoch": 40.18045112781955, + "grad_norm": 0.028405655175447464, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26720 + }, + { + "epoch": 40.19548872180451, + "grad_norm": 0.016238069161772728, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26730 + }, + { + "epoch": 40.21052631578947, + "grad_norm": 0.02470601163804531, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26740 + }, + { + "epoch": 40.225563909774436, + "grad_norm": 0.020551232621073723, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 26750 + }, + { + "epoch": 40.2406015037594, + "grad_norm": 0.028799623250961304, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26760 + }, + { + "epoch": 40.255639097744364, + "grad_norm": 0.022265110164880753, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26770 + }, + { + "epoch": 40.27067669172932, + "grad_norm": 0.028046155348420143, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 26780 + }, + { + "epoch": 40.285714285714285, + "grad_norm": 0.02376970462501049, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26790 + }, + { + "epoch": 40.30075187969925, + "grad_norm": 0.02284211479127407, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26800 + }, + { + "epoch": 40.31578947368421, + "grad_norm": 0.030756574124097824, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 26810 + }, + { + "epoch": 40.330827067669176, + "grad_norm": 0.021828658878803253, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26820 + }, + { + "epoch": 40.34586466165413, + "grad_norm": 0.02741621434688568, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 26830 + }, + { + "epoch": 40.3609022556391, + "grad_norm": 0.04032377153635025, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26840 + }, + { + "epoch": 40.37593984962406, + "grad_norm": 0.04235369712114334, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26850 + }, + { + "epoch": 40.390977443609025, + "grad_norm": 0.033914994448423386, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26860 + }, + { + "epoch": 40.40601503759399, + "grad_norm": 0.026411259546875954, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26870 + }, + { + "epoch": 40.421052631578945, + "grad_norm": 0.034823939204216, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 26880 + }, + { + "epoch": 40.43609022556391, + "grad_norm": 0.03633272647857666, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 26890 + }, + { + "epoch": 40.45112781954887, + "grad_norm": 0.026859251782298088, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26900 + }, + { + "epoch": 40.46616541353384, + "grad_norm": 0.029268495738506317, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 26910 + }, + { + "epoch": 40.4812030075188, + "grad_norm": 0.03576697036623955, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 26920 + }, + { + "epoch": 40.49624060150376, + "grad_norm": 0.035132717341184616, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 26930 + }, + { + "epoch": 40.51127819548872, + "grad_norm": 0.03250621259212494, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 26940 + }, + { + "epoch": 40.526315789473685, + "grad_norm": 0.03211355581879616, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26950 + }, + { + "epoch": 40.54135338345865, + "grad_norm": 0.03954680636525154, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26960 + }, + { + "epoch": 40.556390977443606, + "grad_norm": 0.0318928137421608, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 26970 + }, + { + "epoch": 40.57142857142857, + "grad_norm": 0.02574048936367035, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 26980 + }, + { + "epoch": 40.58646616541353, + "grad_norm": 0.02659679763019085, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 26990 + }, + { + "epoch": 40.6015037593985, + "grad_norm": 0.047241002321243286, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27000 + }, + { + "epoch": 40.6015037593985, + "eval_cer": 0.019376692476868067, + "eval_loss": 0.096665158867836, + "eval_runtime": 158.4268, + "eval_samples_per_second": 101.163, + "eval_steps_per_second": 0.795, + "eval_wer": 0.06886522956642674, + "step": 27000 + }, + { + "epoch": 40.61654135338346, + "grad_norm": 0.024685131385922432, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27010 + }, + { + "epoch": 40.63157894736842, + "grad_norm": 0.03946291655302048, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27020 + }, + { + "epoch": 40.64661654135338, + "grad_norm": 0.03703926503658295, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27030 + }, + { + "epoch": 40.661654135338345, + "grad_norm": 0.03361457586288452, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 27040 + }, + { + "epoch": 40.67669172932331, + "grad_norm": 0.034435611218214035, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 27050 + }, + { + "epoch": 40.69172932330827, + "grad_norm": 0.028128262609243393, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27060 + }, + { + "epoch": 40.70676691729323, + "grad_norm": 0.0233576949685812, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27070 + }, + { + "epoch": 40.721804511278194, + "grad_norm": 0.029901275411248207, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27080 + }, + { + "epoch": 40.73684210526316, + "grad_norm": 0.027990905568003654, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 27090 + }, + { + "epoch": 40.75187969924812, + "grad_norm": 0.024874990805983543, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27100 + }, + { + "epoch": 40.766917293233085, + "grad_norm": 0.050989944487810135, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27110 + }, + { + "epoch": 40.78195488721804, + "grad_norm": 0.03802023082971573, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27120 + }, + { + "epoch": 40.796992481203006, + "grad_norm": 0.02711162529885769, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27130 + }, + { + "epoch": 40.81203007518797, + "grad_norm": 0.025526318699121475, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27140 + }, + { + "epoch": 40.82706766917293, + "grad_norm": 0.022352254018187523, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 27150 + }, + { + "epoch": 40.8421052631579, + "grad_norm": 0.03126024454832077, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27160 + }, + { + "epoch": 40.857142857142854, + "grad_norm": 0.0294490996748209, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27170 + }, + { + "epoch": 40.87218045112782, + "grad_norm": 0.019903957843780518, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27180 + }, + { + "epoch": 40.88721804511278, + "grad_norm": 0.030285237357020378, + "learning_rate": 0.0005, + "loss": 0.0024, + "step": 27190 + }, + { + "epoch": 40.902255639097746, + "grad_norm": 0.03863034397363663, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 27200 + }, + { + "epoch": 40.91729323308271, + "grad_norm": 0.022112032398581505, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27210 + }, + { + "epoch": 40.932330827067666, + "grad_norm": 0.02361234463751316, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27220 + }, + { + "epoch": 40.94736842105263, + "grad_norm": 0.04299623891711235, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27230 + }, + { + "epoch": 40.962406015037594, + "grad_norm": 0.030792804434895515, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27240 + }, + { + "epoch": 40.97744360902256, + "grad_norm": 0.035529449582099915, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27250 + }, + { + "epoch": 40.99248120300752, + "grad_norm": 0.047928549349308014, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27260 + }, + { + "epoch": 41.00751879699248, + "grad_norm": 0.019746774807572365, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27270 + }, + { + "epoch": 41.02255639097744, + "grad_norm": 0.029614031314849854, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27280 + }, + { + "epoch": 41.037593984962406, + "grad_norm": 0.03267286717891693, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27290 + }, + { + "epoch": 41.05263157894737, + "grad_norm": 0.015097795985639095, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 27300 + }, + { + "epoch": 41.067669172932334, + "grad_norm": 0.038016267120838165, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27310 + }, + { + "epoch": 41.08270676691729, + "grad_norm": 0.027099402621388435, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27320 + }, + { + "epoch": 41.097744360902254, + "grad_norm": 0.016774015501141548, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27330 + }, + { + "epoch": 41.11278195488722, + "grad_norm": 0.02944386936724186, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27340 + }, + { + "epoch": 41.12781954887218, + "grad_norm": 0.02072877250611782, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27350 + }, + { + "epoch": 41.142857142857146, + "grad_norm": 0.023544490337371826, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27360 + }, + { + "epoch": 41.1578947368421, + "grad_norm": 0.017509208992123604, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 27370 + }, + { + "epoch": 41.17293233082707, + "grad_norm": 0.03746584430336952, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27380 + }, + { + "epoch": 41.18796992481203, + "grad_norm": 0.01842476986348629, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 27390 + }, + { + "epoch": 41.203007518796994, + "grad_norm": 0.02917494811117649, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27400 + }, + { + "epoch": 41.21804511278196, + "grad_norm": 0.029312320053577423, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27410 + }, + { + "epoch": 41.233082706766915, + "grad_norm": 0.04194508492946625, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27420 + }, + { + "epoch": 41.24812030075188, + "grad_norm": 0.039952877908945084, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27430 + }, + { + "epoch": 41.26315789473684, + "grad_norm": 0.03502289578318596, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27440 + }, + { + "epoch": 41.278195488721806, + "grad_norm": 0.02627694047987461, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27450 + }, + { + "epoch": 41.29323308270677, + "grad_norm": 0.04065695405006409, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27460 + }, + { + "epoch": 41.30827067669173, + "grad_norm": 0.022674674168229103, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27470 + }, + { + "epoch": 41.32330827067669, + "grad_norm": 0.042893774807453156, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27480 + }, + { + "epoch": 41.338345864661655, + "grad_norm": 0.034731995314359665, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 27490 + }, + { + "epoch": 41.35338345864662, + "grad_norm": 0.025748996064066887, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27500 + }, + { + "epoch": 41.36842105263158, + "grad_norm": 0.024762693792581558, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27510 + }, + { + "epoch": 41.38345864661654, + "grad_norm": 0.020467650145292282, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 27520 + }, + { + "epoch": 41.3984962406015, + "grad_norm": 0.031059004366397858, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27530 + }, + { + "epoch": 41.41353383458647, + "grad_norm": 0.022822469472885132, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27540 + }, + { + "epoch": 41.42857142857143, + "grad_norm": 0.02401767112314701, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27550 + }, + { + "epoch": 41.443609022556394, + "grad_norm": 0.03552922233939171, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27560 + }, + { + "epoch": 41.45864661654135, + "grad_norm": 0.030857088044285774, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27570 + }, + { + "epoch": 41.473684210526315, + "grad_norm": 0.03771974891424179, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27580 + }, + { + "epoch": 41.48872180451128, + "grad_norm": 0.023044085130095482, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27590 + }, + { + "epoch": 41.50375939849624, + "grad_norm": 0.02118164300918579, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27600 + }, + { + "epoch": 41.5187969924812, + "grad_norm": 0.029012657701969147, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27610 + }, + { + "epoch": 41.53383458646616, + "grad_norm": 0.02446054294705391, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27620 + }, + { + "epoch": 41.54887218045113, + "grad_norm": 0.033321645110845566, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27630 + }, + { + "epoch": 41.56390977443609, + "grad_norm": 0.02636510320007801, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27640 + }, + { + "epoch": 41.578947368421055, + "grad_norm": 0.027779126539826393, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 27650 + }, + { + "epoch": 41.59398496240601, + "grad_norm": 0.017922641709446907, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 27660 + }, + { + "epoch": 41.609022556390975, + "grad_norm": 0.027365058660507202, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27670 + }, + { + "epoch": 41.62406015037594, + "grad_norm": 0.0455225370824337, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27680 + }, + { + "epoch": 41.6390977443609, + "grad_norm": 0.04712602496147156, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 27690 + }, + { + "epoch": 41.65413533834587, + "grad_norm": 0.03551556169986725, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27700 + }, + { + "epoch": 41.669172932330824, + "grad_norm": 0.027157053351402283, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27710 + }, + { + "epoch": 41.68421052631579, + "grad_norm": 0.031927239149808884, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27720 + }, + { + "epoch": 41.69924812030075, + "grad_norm": 0.019553055986762047, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27730 + }, + { + "epoch": 41.714285714285715, + "grad_norm": 0.02712133526802063, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 27740 + }, + { + "epoch": 41.72932330827068, + "grad_norm": 0.02005056105554104, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 27750 + }, + { + "epoch": 41.744360902255636, + "grad_norm": 0.02942042611539364, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 27760 + }, + { + "epoch": 41.7593984962406, + "grad_norm": 0.021262338384985924, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 27770 + }, + { + "epoch": 41.774436090225564, + "grad_norm": 0.03387856110930443, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27780 + }, + { + "epoch": 41.78947368421053, + "grad_norm": 0.03159164637327194, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27790 + }, + { + "epoch": 41.80451127819549, + "grad_norm": 0.015025395900011063, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 27800 + }, + { + "epoch": 41.81954887218045, + "grad_norm": 0.016219964250922203, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27810 + }, + { + "epoch": 41.83458646616541, + "grad_norm": 0.02565586008131504, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27820 + }, + { + "epoch": 41.849624060150376, + "grad_norm": 0.03824986517429352, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 27830 + }, + { + "epoch": 41.86466165413534, + "grad_norm": 0.024678746238350868, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27840 + }, + { + "epoch": 41.8796992481203, + "grad_norm": 0.018768854439258575, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27850 + }, + { + "epoch": 41.89473684210526, + "grad_norm": 0.019597947597503662, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27860 + }, + { + "epoch": 41.909774436090224, + "grad_norm": 0.014974048361182213, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27870 + }, + { + "epoch": 41.92481203007519, + "grad_norm": 0.023916922509670258, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27880 + }, + { + "epoch": 41.93984962406015, + "grad_norm": 0.03140264376997948, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 27890 + }, + { + "epoch": 41.954887218045116, + "grad_norm": 0.043228428810834885, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27900 + }, + { + "epoch": 41.96992481203007, + "grad_norm": 0.026687616482377052, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 27910 + }, + { + "epoch": 41.984962406015036, + "grad_norm": 0.023808272555470467, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 27920 + }, + { + "epoch": 42.0, + "grad_norm": 0.04353948310017586, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 27930 + }, + { + "epoch": 42.015037593984964, + "grad_norm": 0.02247891016304493, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27940 + }, + { + "epoch": 42.03007518796993, + "grad_norm": 0.0345262810587883, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27950 + }, + { + "epoch": 42.045112781954884, + "grad_norm": 0.02756444737315178, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 27960 + }, + { + "epoch": 42.06015037593985, + "grad_norm": 0.03184790536761284, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 27970 + }, + { + "epoch": 42.07518796992481, + "grad_norm": 0.028907712548971176, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27980 + }, + { + "epoch": 42.090225563909776, + "grad_norm": 0.015311412513256073, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 27990 + }, + { + "epoch": 42.10526315789474, + "grad_norm": 0.02954692393541336, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 28000 + }, + { + "epoch": 42.10526315789474, + "eval_cer": 0.01837565743928165, + "eval_loss": 0.097585029900074, + "eval_runtime": 161.1173, + "eval_samples_per_second": 99.474, + "eval_steps_per_second": 0.782, + "eval_wer": 0.06713150031009792, + "step": 28000 + }, + { + "epoch": 42.1203007518797, + "grad_norm": 0.013470249250531197, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28010 + }, + { + "epoch": 42.13533834586466, + "grad_norm": 0.025999337434768677, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28020 + }, + { + "epoch": 42.150375939849624, + "grad_norm": 0.023759393021464348, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28030 + }, + { + "epoch": 42.16541353383459, + "grad_norm": 0.026712458580732346, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28040 + }, + { + "epoch": 42.18045112781955, + "grad_norm": 0.021211965009570122, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28050 + }, + { + "epoch": 42.19548872180451, + "grad_norm": 0.024399466812610626, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 28060 + }, + { + "epoch": 42.21052631578947, + "grad_norm": 0.03514518961310387, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28070 + }, + { + "epoch": 42.225563909774436, + "grad_norm": 0.022147828713059425, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 28080 + }, + { + "epoch": 42.2406015037594, + "grad_norm": 0.031161481514573097, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28090 + }, + { + "epoch": 42.255639097744364, + "grad_norm": 0.04453691840171814, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 28100 + }, + { + "epoch": 42.27067669172932, + "grad_norm": 0.02615354023873806, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28110 + }, + { + "epoch": 42.285714285714285, + "grad_norm": 0.027012750506401062, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 28120 + }, + { + "epoch": 42.30075187969925, + "grad_norm": 0.03117816522717476, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28130 + }, + { + "epoch": 42.31578947368421, + "grad_norm": 0.016094405204057693, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28140 + }, + { + "epoch": 42.330827067669176, + "grad_norm": 0.02764665149152279, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 28150 + }, + { + "epoch": 42.34586466165413, + "grad_norm": 0.018073156476020813, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28160 + }, + { + "epoch": 42.3609022556391, + "grad_norm": 0.02992287650704384, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28170 + }, + { + "epoch": 42.37593984962406, + "grad_norm": 0.016769621521234512, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28180 + }, + { + "epoch": 42.390977443609025, + "grad_norm": 0.03150714933872223, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28190 + }, + { + "epoch": 42.40601503759399, + "grad_norm": 0.02754213847219944, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28200 + }, + { + "epoch": 42.421052631578945, + "grad_norm": 0.03097599186003208, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28210 + }, + { + "epoch": 42.43609022556391, + "grad_norm": 0.034128475934267044, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28220 + }, + { + "epoch": 42.45112781954887, + "grad_norm": 0.05247306451201439, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28230 + }, + { + "epoch": 42.46616541353384, + "grad_norm": 0.015239577740430832, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28240 + }, + { + "epoch": 42.4812030075188, + "grad_norm": 0.027879787608981133, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28250 + }, + { + "epoch": 42.49624060150376, + "grad_norm": 0.03300035744905472, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28260 + }, + { + "epoch": 42.51127819548872, + "grad_norm": 0.03826013207435608, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28270 + }, + { + "epoch": 42.526315789473685, + "grad_norm": 0.02038433589041233, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28280 + }, + { + "epoch": 42.54135338345865, + "grad_norm": 0.03302670270204544, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28290 + }, + { + "epoch": 42.556390977443606, + "grad_norm": 0.018936343491077423, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28300 + }, + { + "epoch": 42.57142857142857, + "grad_norm": 0.024404877796769142, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28310 + }, + { + "epoch": 42.58646616541353, + "grad_norm": 0.02551860176026821, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28320 + }, + { + "epoch": 42.6015037593985, + "grad_norm": 0.0309690460562706, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 28330 + }, + { + "epoch": 42.61654135338346, + "grad_norm": 0.023480910807847977, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28340 + }, + { + "epoch": 42.63157894736842, + "grad_norm": 0.03182827681303024, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 28350 + }, + { + "epoch": 42.64661654135338, + "grad_norm": 0.03461227938532829, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28360 + }, + { + "epoch": 42.661654135338345, + "grad_norm": 0.022408179938793182, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28370 + }, + { + "epoch": 42.67669172932331, + "grad_norm": 0.03788776323199272, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28380 + }, + { + "epoch": 42.69172932330827, + "grad_norm": 0.034337498247623444, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 28390 + }, + { + "epoch": 42.70676691729323, + "grad_norm": 0.023281654343008995, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28400 + }, + { + "epoch": 42.721804511278194, + "grad_norm": 0.028789425268769264, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28410 + }, + { + "epoch": 42.73684210526316, + "grad_norm": 0.02680501714348793, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 28420 + }, + { + "epoch": 42.75187969924812, + "grad_norm": 0.02231026440858841, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 28430 + }, + { + "epoch": 42.766917293233085, + "grad_norm": 0.05130762979388237, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 28440 + }, + { + "epoch": 42.78195488721804, + "grad_norm": 0.023621194064617157, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28450 + }, + { + "epoch": 42.796992481203006, + "grad_norm": 0.029675913974642754, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28460 + }, + { + "epoch": 42.81203007518797, + "grad_norm": 0.033151499927043915, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28470 + }, + { + "epoch": 42.82706766917293, + "grad_norm": 0.019587047398090363, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28480 + }, + { + "epoch": 42.8421052631579, + "grad_norm": 0.03666991740465164, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 28490 + }, + { + "epoch": 42.857142857142854, + "grad_norm": 0.037990398705005646, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28500 + }, + { + "epoch": 42.87218045112782, + "grad_norm": 0.032552894204854965, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28510 + }, + { + "epoch": 42.88721804511278, + "grad_norm": 0.03454093635082245, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28520 + }, + { + "epoch": 42.902255639097746, + "grad_norm": 0.024828361347317696, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28530 + }, + { + "epoch": 42.91729323308271, + "grad_norm": 0.026338757947087288, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28540 + }, + { + "epoch": 42.932330827067666, + "grad_norm": 0.035806991159915924, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28550 + }, + { + "epoch": 42.94736842105263, + "grad_norm": 0.05872581899166107, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 28560 + }, + { + "epoch": 42.962406015037594, + "grad_norm": 0.04030924290418625, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28570 + }, + { + "epoch": 42.97744360902256, + "grad_norm": 0.018687549978494644, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 28580 + }, + { + "epoch": 42.99248120300752, + "grad_norm": 0.022393453866243362, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28590 + }, + { + "epoch": 43.00751879699248, + "grad_norm": 0.028223799541592598, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28600 + }, + { + "epoch": 43.02255639097744, + "grad_norm": 0.018043071031570435, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28610 + }, + { + "epoch": 43.037593984962406, + "grad_norm": 0.015348775312304497, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 28620 + }, + { + "epoch": 43.05263157894737, + "grad_norm": 0.032419852912425995, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 28630 + }, + { + "epoch": 43.067669172932334, + "grad_norm": 0.03337999805808067, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28640 + }, + { + "epoch": 43.08270676691729, + "grad_norm": 0.035753585398197174, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28650 + }, + { + "epoch": 43.097744360902254, + "grad_norm": 0.022903714329004288, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28660 + }, + { + "epoch": 43.11278195488722, + "grad_norm": 0.037555307149887085, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28670 + }, + { + "epoch": 43.12781954887218, + "grad_norm": 0.045158546417951584, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28680 + }, + { + "epoch": 43.142857142857146, + "grad_norm": 0.03091997280716896, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28690 + }, + { + "epoch": 43.1578947368421, + "grad_norm": 0.018241193145513535, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28700 + }, + { + "epoch": 43.17293233082707, + "grad_norm": 0.01591287925839424, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28710 + }, + { + "epoch": 43.18796992481203, + "grad_norm": 0.023748666048049927, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28720 + }, + { + "epoch": 43.203007518796994, + "grad_norm": 0.024053890258073807, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28730 + }, + { + "epoch": 43.21804511278196, + "grad_norm": 0.027382465079426765, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28740 + }, + { + "epoch": 43.233082706766915, + "grad_norm": 0.025523612275719643, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28750 + }, + { + "epoch": 43.24812030075188, + "grad_norm": 0.028210218995809555, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28760 + }, + { + "epoch": 43.26315789473684, + "grad_norm": 0.029698552563786507, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28770 + }, + { + "epoch": 43.278195488721806, + "grad_norm": 0.03362550586462021, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28780 + }, + { + "epoch": 43.29323308270677, + "grad_norm": 0.02329862490296364, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28790 + }, + { + "epoch": 43.30827067669173, + "grad_norm": 0.018285434693098068, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28800 + }, + { + "epoch": 43.32330827067669, + "grad_norm": 0.027988288551568985, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28810 + }, + { + "epoch": 43.338345864661655, + "grad_norm": 0.027439571917057037, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28820 + }, + { + "epoch": 43.35338345864662, + "grad_norm": 0.025220049545168877, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28830 + }, + { + "epoch": 43.36842105263158, + "grad_norm": 0.030995750799775124, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28840 + }, + { + "epoch": 43.38345864661654, + "grad_norm": 0.025441573932766914, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28850 + }, + { + "epoch": 43.3984962406015, + "grad_norm": 0.022687841206789017, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28860 + }, + { + "epoch": 43.41353383458647, + "grad_norm": 0.028760509565472603, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28870 + }, + { + "epoch": 43.42857142857143, + "grad_norm": 0.02057790197432041, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28880 + }, + { + "epoch": 43.443609022556394, + "grad_norm": 0.04362573102116585, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28890 + }, + { + "epoch": 43.45864661654135, + "grad_norm": 0.03351295739412308, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28900 + }, + { + "epoch": 43.473684210526315, + "grad_norm": 0.03902207687497139, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 28910 + }, + { + "epoch": 43.48872180451128, + "grad_norm": 0.032186657190322876, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28920 + }, + { + "epoch": 43.50375939849624, + "grad_norm": 0.024735093116760254, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 28930 + }, + { + "epoch": 43.5187969924812, + "grad_norm": 0.018030604347586632, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 28940 + }, + { + "epoch": 43.53383458646616, + "grad_norm": 0.025249898433685303, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 28950 + }, + { + "epoch": 43.54887218045113, + "grad_norm": 0.035777851939201355, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28960 + }, + { + "epoch": 43.56390977443609, + "grad_norm": 0.029922226443886757, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 28970 + }, + { + "epoch": 43.578947368421055, + "grad_norm": 0.0236323494464159, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 28980 + }, + { + "epoch": 43.59398496240601, + "grad_norm": 0.015603124164044857, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 28990 + }, + { + "epoch": 43.609022556390975, + "grad_norm": 0.044788703322410583, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29000 + }, + { + "epoch": 43.609022556390975, + "eval_cer": 0.018601434363519888, + "eval_loss": 0.09733881056308746, + "eval_runtime": 159.7072, + "eval_samples_per_second": 100.352, + "eval_steps_per_second": 0.789, + "eval_wer": 0.06759194873047793, + "step": 29000 + }, + { + "epoch": 43.62406015037594, + "grad_norm": 0.032682377845048904, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29010 + }, + { + "epoch": 43.6390977443609, + "grad_norm": 0.03307877853512764, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29020 + }, + { + "epoch": 43.65413533834587, + "grad_norm": 0.026527438312768936, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29030 + }, + { + "epoch": 43.669172932330824, + "grad_norm": 0.02744593285024166, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29040 + }, + { + "epoch": 43.68421052631579, + "grad_norm": 0.029821153730154037, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29050 + }, + { + "epoch": 43.69924812030075, + "grad_norm": 0.03640605881810188, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29060 + }, + { + "epoch": 43.714285714285715, + "grad_norm": 0.038532938808202744, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29070 + }, + { + "epoch": 43.72932330827068, + "grad_norm": 0.020721426233649254, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29080 + }, + { + "epoch": 43.744360902255636, + "grad_norm": 0.029370836913585663, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 29090 + }, + { + "epoch": 43.7593984962406, + "grad_norm": 0.026895351707935333, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29100 + }, + { + "epoch": 43.774436090225564, + "grad_norm": 0.03253600373864174, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29110 + }, + { + "epoch": 43.78947368421053, + "grad_norm": 0.021299535408616066, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29120 + }, + { + "epoch": 43.80451127819549, + "grad_norm": 0.025056906044483185, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29130 + }, + { + "epoch": 43.81954887218045, + "grad_norm": 0.025626858696341515, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29140 + }, + { + "epoch": 43.83458646616541, + "grad_norm": 0.025241168215870857, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29150 + }, + { + "epoch": 43.849624060150376, + "grad_norm": 0.021622365340590477, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29160 + }, + { + "epoch": 43.86466165413534, + "grad_norm": 0.020686523988842964, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29170 + }, + { + "epoch": 43.8796992481203, + "grad_norm": 0.02833201363682747, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29180 + }, + { + "epoch": 43.89473684210526, + "grad_norm": 0.029819289222359657, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29190 + }, + { + "epoch": 43.909774436090224, + "grad_norm": 0.0343354158103466, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29200 + }, + { + "epoch": 43.92481203007519, + "grad_norm": 0.02716091461479664, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29210 + }, + { + "epoch": 43.93984962406015, + "grad_norm": 0.03213988617062569, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29220 + }, + { + "epoch": 43.954887218045116, + "grad_norm": 0.023171283304691315, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29230 + }, + { + "epoch": 43.96992481203007, + "grad_norm": 0.021117648109793663, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29240 + }, + { + "epoch": 43.984962406015036, + "grad_norm": 0.0240913238376379, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 29250 + }, + { + "epoch": 44.0, + "grad_norm": 0.019245855510234833, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 29260 + }, + { + "epoch": 44.015037593984964, + "grad_norm": 0.038796231150627136, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29270 + }, + { + "epoch": 44.03007518796993, + "grad_norm": 0.01789679005742073, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29280 + }, + { + "epoch": 44.045112781954884, + "grad_norm": 0.027316641062498093, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29290 + }, + { + "epoch": 44.06015037593985, + "grad_norm": 0.03199198096990585, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29300 + }, + { + "epoch": 44.07518796992481, + "grad_norm": 0.031416986137628555, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 29310 + }, + { + "epoch": 44.090225563909776, + "grad_norm": 0.04552570730447769, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29320 + }, + { + "epoch": 44.10526315789474, + "grad_norm": 0.043427370488643646, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29330 + }, + { + "epoch": 44.1203007518797, + "grad_norm": 0.023930173367261887, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 29340 + }, + { + "epoch": 44.13533834586466, + "grad_norm": 0.03000332973897457, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 29350 + }, + { + "epoch": 44.150375939849624, + "grad_norm": 0.03329074755311012, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29360 + }, + { + "epoch": 44.16541353383459, + "grad_norm": 0.02192349173128605, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29370 + }, + { + "epoch": 44.18045112781955, + "grad_norm": 0.023860350251197815, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29380 + }, + { + "epoch": 44.19548872180451, + "grad_norm": 0.02227587252855301, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29390 + }, + { + "epoch": 44.21052631578947, + "grad_norm": 0.04083609580993652, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29400 + }, + { + "epoch": 44.225563909774436, + "grad_norm": 0.031337808817625046, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29410 + }, + { + "epoch": 44.2406015037594, + "grad_norm": 0.046246662735939026, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29420 + }, + { + "epoch": 44.255639097744364, + "grad_norm": 0.01692219264805317, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29430 + }, + { + "epoch": 44.27067669172932, + "grad_norm": 0.024403661489486694, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29440 + }, + { + "epoch": 44.285714285714285, + "grad_norm": 0.014682470820844173, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29450 + }, + { + "epoch": 44.30075187969925, + "grad_norm": 0.03339656814932823, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29460 + }, + { + "epoch": 44.31578947368421, + "grad_norm": 0.03890642151236534, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29470 + }, + { + "epoch": 44.330827067669176, + "grad_norm": 0.021977214142680168, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29480 + }, + { + "epoch": 44.34586466165413, + "grad_norm": 0.04402611404657364, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29490 + }, + { + "epoch": 44.3609022556391, + "grad_norm": 0.054150618612766266, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29500 + }, + { + "epoch": 44.37593984962406, + "grad_norm": 0.018683796748518944, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29510 + }, + { + "epoch": 44.390977443609025, + "grad_norm": 0.02765027992427349, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 29520 + }, + { + "epoch": 44.40601503759399, + "grad_norm": 0.03544798865914345, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29530 + }, + { + "epoch": 44.421052631578945, + "grad_norm": 0.020323792472481728, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29540 + }, + { + "epoch": 44.43609022556391, + "grad_norm": 0.03211567923426628, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29550 + }, + { + "epoch": 44.45112781954887, + "grad_norm": 0.026003848761320114, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29560 + }, + { + "epoch": 44.46616541353384, + "grad_norm": 0.037049632519483566, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 29570 + }, + { + "epoch": 44.4812030075188, + "grad_norm": 0.024818407371640205, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29580 + }, + { + "epoch": 44.49624060150376, + "grad_norm": 0.01864086464047432, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29590 + }, + { + "epoch": 44.51127819548872, + "grad_norm": 0.030035611242055893, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29600 + }, + { + "epoch": 44.526315789473685, + "grad_norm": 0.026362327858805656, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29610 + }, + { + "epoch": 44.54135338345865, + "grad_norm": 0.031159818172454834, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29620 + }, + { + "epoch": 44.556390977443606, + "grad_norm": 0.03309270367026329, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29630 + }, + { + "epoch": 44.57142857142857, + "grad_norm": 0.02795824222266674, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29640 + }, + { + "epoch": 44.58646616541353, + "grad_norm": 0.03245827183127403, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29650 + }, + { + "epoch": 44.6015037593985, + "grad_norm": 0.02943698689341545, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29660 + }, + { + "epoch": 44.61654135338346, + "grad_norm": 0.022860467433929443, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 29670 + }, + { + "epoch": 44.63157894736842, + "grad_norm": 0.026967622339725494, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29680 + }, + { + "epoch": 44.64661654135338, + "grad_norm": 0.022923510521650314, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29690 + }, + { + "epoch": 44.661654135338345, + "grad_norm": 0.02575310878455639, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29700 + }, + { + "epoch": 44.67669172932331, + "grad_norm": 0.02108384110033512, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29710 + }, + { + "epoch": 44.69172932330827, + "grad_norm": 0.03390626609325409, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29720 + }, + { + "epoch": 44.70676691729323, + "grad_norm": 0.02801351435482502, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 29730 + }, + { + "epoch": 44.721804511278194, + "grad_norm": 0.02753012254834175, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29740 + }, + { + "epoch": 44.73684210526316, + "grad_norm": 0.024656828492879868, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29750 + }, + { + "epoch": 44.75187969924812, + "grad_norm": 0.03077618032693863, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29760 + }, + { + "epoch": 44.766917293233085, + "grad_norm": 0.02936864085495472, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 29770 + }, + { + "epoch": 44.78195488721804, + "grad_norm": 0.03729255869984627, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 29780 + }, + { + "epoch": 44.796992481203006, + "grad_norm": 0.03485661372542381, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 29790 + }, + { + "epoch": 44.81203007518797, + "grad_norm": 0.03531154617667198, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 29800 + }, + { + "epoch": 44.82706766917293, + "grad_norm": 0.0183413103222847, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29810 + }, + { + "epoch": 44.8421052631579, + "grad_norm": 0.025420991703867912, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29820 + }, + { + "epoch": 44.857142857142854, + "grad_norm": 0.02033226564526558, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29830 + }, + { + "epoch": 44.87218045112782, + "grad_norm": 0.02576085552573204, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29840 + }, + { + "epoch": 44.88721804511278, + "grad_norm": 0.013679787516593933, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 29850 + }, + { + "epoch": 44.902255639097746, + "grad_norm": 0.01814301684498787, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29860 + }, + { + "epoch": 44.91729323308271, + "grad_norm": 0.02429485134780407, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29870 + }, + { + "epoch": 44.932330827067666, + "grad_norm": 0.031085088849067688, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 29880 + }, + { + "epoch": 44.94736842105263, + "grad_norm": 0.028634807094931602, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 29890 + }, + { + "epoch": 44.962406015037594, + "grad_norm": 0.031280066817998886, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 29900 + }, + { + "epoch": 44.97744360902256, + "grad_norm": 0.0451592318713665, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29910 + }, + { + "epoch": 44.99248120300752, + "grad_norm": 0.030531885102391243, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29920 + }, + { + "epoch": 45.00751879699248, + "grad_norm": 0.013329938985407352, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 29930 + }, + { + "epoch": 45.02255639097744, + "grad_norm": 0.023131513968110085, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29940 + }, + { + "epoch": 45.037593984962406, + "grad_norm": 0.02537067048251629, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29950 + }, + { + "epoch": 45.05263157894737, + "grad_norm": 0.028196299448609352, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29960 + }, + { + "epoch": 45.067669172932334, + "grad_norm": 0.02631358988583088, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29970 + }, + { + "epoch": 45.08270676691729, + "grad_norm": 0.024056050926446915, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 29980 + }, + { + "epoch": 45.097744360902254, + "grad_norm": 0.030398845672607422, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 29990 + }, + { + "epoch": 45.11278195488722, + "grad_norm": 0.028872868046164513, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30000 + }, + { + "epoch": 45.11278195488722, + "eval_cer": 0.019341329826083765, + "eval_loss": 0.095069020986557, + "eval_runtime": 162.7381, + "eval_samples_per_second": 98.483, + "eval_steps_per_second": 0.774, + "eval_wer": 0.06683549775413934, + "step": 30000 + }, + { + "epoch": 45.12781954887218, + "grad_norm": 0.026854632422327995, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30010 + }, + { + "epoch": 45.142857142857146, + "grad_norm": 0.0261649489402771, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 30020 + }, + { + "epoch": 45.1578947368421, + "grad_norm": 0.023663559928536415, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30030 + }, + { + "epoch": 45.17293233082707, + "grad_norm": 0.029630402103066444, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30040 + }, + { + "epoch": 45.18796992481203, + "grad_norm": 0.026158249005675316, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30050 + }, + { + "epoch": 45.203007518796994, + "grad_norm": 0.029941854998469353, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30060 + }, + { + "epoch": 45.21804511278196, + "grad_norm": 0.021942973136901855, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30070 + }, + { + "epoch": 45.233082706766915, + "grad_norm": 0.025299210101366043, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30080 + }, + { + "epoch": 45.24812030075188, + "grad_norm": 0.024404868483543396, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30090 + }, + { + "epoch": 45.26315789473684, + "grad_norm": 0.05113166198134422, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30100 + }, + { + "epoch": 45.278195488721806, + "grad_norm": 0.019192799925804138, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30110 + }, + { + "epoch": 45.29323308270677, + "grad_norm": 0.0162139143794775, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30120 + }, + { + "epoch": 45.30827067669173, + "grad_norm": 0.02834874950349331, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30130 + }, + { + "epoch": 45.32330827067669, + "grad_norm": 0.03377317264676094, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30140 + }, + { + "epoch": 45.338345864661655, + "grad_norm": 0.0361856147646904, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30150 + }, + { + "epoch": 45.35338345864662, + "grad_norm": 0.02440139465034008, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30160 + }, + { + "epoch": 45.36842105263158, + "grad_norm": 0.013842448592185974, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30170 + }, + { + "epoch": 45.38345864661654, + "grad_norm": 0.025715786963701248, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30180 + }, + { + "epoch": 45.3984962406015, + "grad_norm": 0.020905816927552223, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 30190 + }, + { + "epoch": 45.41353383458647, + "grad_norm": 0.016318626701831818, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 30200 + }, + { + "epoch": 45.42857142857143, + "grad_norm": 0.02089015766978264, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30210 + }, + { + "epoch": 45.443609022556394, + "grad_norm": 0.019128436222672462, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30220 + }, + { + "epoch": 45.45864661654135, + "grad_norm": 0.027089159935712814, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30230 + }, + { + "epoch": 45.473684210526315, + "grad_norm": 0.029036035761237144, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30240 + }, + { + "epoch": 45.48872180451128, + "grad_norm": 0.021726571023464203, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30250 + }, + { + "epoch": 45.50375939849624, + "grad_norm": 0.03917199373245239, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30260 + }, + { + "epoch": 45.5187969924812, + "grad_norm": 0.03513648360967636, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30270 + }, + { + "epoch": 45.53383458646616, + "grad_norm": 0.031827569007873535, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30280 + }, + { + "epoch": 45.54887218045113, + "grad_norm": 0.021343229338526726, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30290 + }, + { + "epoch": 45.56390977443609, + "grad_norm": 0.030830049887299538, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30300 + }, + { + "epoch": 45.578947368421055, + "grad_norm": 0.027749786153435707, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30310 + }, + { + "epoch": 45.59398496240601, + "grad_norm": 0.033659279346466064, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 30320 + }, + { + "epoch": 45.609022556390975, + "grad_norm": 0.03372489660978317, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 30330 + }, + { + "epoch": 45.62406015037594, + "grad_norm": 0.03065464273095131, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30340 + }, + { + "epoch": 45.6390977443609, + "grad_norm": 0.02546004019677639, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30350 + }, + { + "epoch": 45.65413533834587, + "grad_norm": 0.03201393783092499, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30360 + }, + { + "epoch": 45.669172932330824, + "grad_norm": 0.016620982438325882, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30370 + }, + { + "epoch": 45.68421052631579, + "grad_norm": 0.015734270215034485, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30380 + }, + { + "epoch": 45.69924812030075, + "grad_norm": 0.025042634457349777, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 30390 + }, + { + "epoch": 45.714285714285715, + "grad_norm": 0.02504882775247097, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30400 + }, + { + "epoch": 45.72932330827068, + "grad_norm": 0.02216869220137596, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30410 + }, + { + "epoch": 45.744360902255636, + "grad_norm": 0.028668759390711784, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30420 + }, + { + "epoch": 45.7593984962406, + "grad_norm": 0.023771734908223152, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30430 + }, + { + "epoch": 45.774436090225564, + "grad_norm": 0.02116999216377735, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30440 + }, + { + "epoch": 45.78947368421053, + "grad_norm": 0.02686053328216076, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 30450 + }, + { + "epoch": 45.80451127819549, + "grad_norm": 0.03297910466790199, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 30460 + }, + { + "epoch": 45.81954887218045, + "grad_norm": 0.03318452462553978, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30470 + }, + { + "epoch": 45.83458646616541, + "grad_norm": 0.02188408374786377, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30480 + }, + { + "epoch": 45.849624060150376, + "grad_norm": 0.042016513645648956, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 30490 + }, + { + "epoch": 45.86466165413534, + "grad_norm": 0.026468006893992424, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30500 + }, + { + "epoch": 45.8796992481203, + "grad_norm": 0.02462773025035858, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30510 + }, + { + "epoch": 45.89473684210526, + "grad_norm": 0.022513601928949356, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30520 + }, + { + "epoch": 45.909774436090224, + "grad_norm": 0.030221566557884216, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 30530 + }, + { + "epoch": 45.92481203007519, + "grad_norm": 0.020578160881996155, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30540 + }, + { + "epoch": 45.93984962406015, + "grad_norm": 0.017466342076659203, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 30550 + }, + { + "epoch": 45.954887218045116, + "grad_norm": 0.034777816385030746, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30560 + }, + { + "epoch": 45.96992481203007, + "grad_norm": 0.021935464814305305, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30570 + }, + { + "epoch": 45.984962406015036, + "grad_norm": 0.026723017916083336, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30580 + }, + { + "epoch": 46.0, + "grad_norm": 0.016592929139733315, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30590 + }, + { + "epoch": 46.015037593984964, + "grad_norm": 0.020083095878362656, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 30600 + }, + { + "epoch": 46.03007518796993, + "grad_norm": 0.025819294154644012, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30610 + }, + { + "epoch": 46.045112781954884, + "grad_norm": 0.03223176300525665, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30620 + }, + { + "epoch": 46.06015037593985, + "grad_norm": 0.03212769702076912, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30630 + }, + { + "epoch": 46.07518796992481, + "grad_norm": 0.03166981786489487, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30640 + }, + { + "epoch": 46.090225563909776, + "grad_norm": 0.03009401075541973, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30650 + }, + { + "epoch": 46.10526315789474, + "grad_norm": 0.01349911279976368, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30660 + }, + { + "epoch": 46.1203007518797, + "grad_norm": 0.024095728993415833, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30670 + }, + { + "epoch": 46.13533834586466, + "grad_norm": 0.028277281671762466, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30680 + }, + { + "epoch": 46.150375939849624, + "grad_norm": 0.020017748698592186, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 30690 + }, + { + "epoch": 46.16541353383459, + "grad_norm": 0.023630401119589806, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30700 + }, + { + "epoch": 46.18045112781955, + "grad_norm": 0.019268784672021866, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 30710 + }, + { + "epoch": 46.19548872180451, + "grad_norm": 0.029977872967720032, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 30720 + }, + { + "epoch": 46.21052631578947, + "grad_norm": 0.026212705299258232, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30730 + }, + { + "epoch": 46.225563909774436, + "grad_norm": 0.025292925536632538, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30740 + }, + { + "epoch": 46.2406015037594, + "grad_norm": 0.020573390647768974, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30750 + }, + { + "epoch": 46.255639097744364, + "grad_norm": 0.022188827395439148, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30760 + }, + { + "epoch": 46.27067669172932, + "grad_norm": 0.025541886687278748, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30770 + }, + { + "epoch": 46.285714285714285, + "grad_norm": 0.01816335879266262, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30780 + }, + { + "epoch": 46.30075187969925, + "grad_norm": 0.023308701813220978, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30790 + }, + { + "epoch": 46.31578947368421, + "grad_norm": 0.028368890285491943, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30800 + }, + { + "epoch": 46.330827067669176, + "grad_norm": 0.020261557772755623, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30810 + }, + { + "epoch": 46.34586466165413, + "grad_norm": 0.019841989502310753, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30820 + }, + { + "epoch": 46.3609022556391, + "grad_norm": 0.026861149817705154, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30830 + }, + { + "epoch": 46.37593984962406, + "grad_norm": 0.02091345191001892, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30840 + }, + { + "epoch": 46.390977443609025, + "grad_norm": 0.020887942984700203, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30850 + }, + { + "epoch": 46.40601503759399, + "grad_norm": 0.026609055697917938, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30860 + }, + { + "epoch": 46.421052631578945, + "grad_norm": 0.017967665567994118, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30870 + }, + { + "epoch": 46.43609022556391, + "grad_norm": 0.028741246089339256, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 30880 + }, + { + "epoch": 46.45112781954887, + "grad_norm": 0.018097998574376106, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 30890 + }, + { + "epoch": 46.46616541353384, + "grad_norm": 0.02708539366722107, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30900 + }, + { + "epoch": 46.4812030075188, + "grad_norm": 0.017268309369683266, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30910 + }, + { + "epoch": 46.49624060150376, + "grad_norm": 0.044214796274900436, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30920 + }, + { + "epoch": 46.51127819548872, + "grad_norm": 0.032014504075050354, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 30930 + }, + { + "epoch": 46.526315789473685, + "grad_norm": 0.016855284571647644, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30940 + }, + { + "epoch": 46.54135338345865, + "grad_norm": 0.03166595846414566, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 30950 + }, + { + "epoch": 46.556390977443606, + "grad_norm": 0.028208842501044273, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30960 + }, + { + "epoch": 46.57142857142857, + "grad_norm": 0.033855680376291275, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 30970 + }, + { + "epoch": 46.58646616541353, + "grad_norm": 0.01727733574807644, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 30980 + }, + { + "epoch": 46.6015037593985, + "grad_norm": 0.04010910913348198, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 30990 + }, + { + "epoch": 46.61654135338346, + "grad_norm": 0.027376657351851463, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 31000 + }, + { + "epoch": 46.61654135338346, + "eval_cer": 0.019308007328229324, + "eval_loss": 0.09365199506282806, + "eval_runtime": 160.3202, + "eval_samples_per_second": 99.969, + "eval_steps_per_second": 0.786, + "eval_wer": 0.06751207502490181, + "step": 31000 + }, + { + "epoch": 46.63157894736842, + "grad_norm": 0.023674001917243004, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31010 + }, + { + "epoch": 46.64661654135338, + "grad_norm": 0.014673136174678802, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31020 + }, + { + "epoch": 46.661654135338345, + "grad_norm": 0.012601487338542938, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31030 + }, + { + "epoch": 46.67669172932331, + "grad_norm": 0.018680140376091003, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31040 + }, + { + "epoch": 46.69172932330827, + "grad_norm": 0.03122525103390217, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 31050 + }, + { + "epoch": 46.70676691729323, + "grad_norm": 0.021286293864250183, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31060 + }, + { + "epoch": 46.721804511278194, + "grad_norm": 0.03236154839396477, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31070 + }, + { + "epoch": 46.73684210526316, + "grad_norm": 0.013495407067239285, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31080 + }, + { + "epoch": 46.75187969924812, + "grad_norm": 0.018509795889258385, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31090 + }, + { + "epoch": 46.766917293233085, + "grad_norm": 0.029791397973895073, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31100 + }, + { + "epoch": 46.78195488721804, + "grad_norm": 0.022087562829256058, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31110 + }, + { + "epoch": 46.796992481203006, + "grad_norm": 0.025269340723752975, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31120 + }, + { + "epoch": 46.81203007518797, + "grad_norm": 0.017589662224054337, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31130 + }, + { + "epoch": 46.82706766917293, + "grad_norm": 0.014708077535033226, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31140 + }, + { + "epoch": 46.8421052631579, + "grad_norm": 0.014012542553246021, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31150 + }, + { + "epoch": 46.857142857142854, + "grad_norm": 0.01734212413430214, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31160 + }, + { + "epoch": 46.87218045112782, + "grad_norm": 0.025004934519529343, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31170 + }, + { + "epoch": 46.88721804511278, + "grad_norm": 0.02571006491780281, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31180 + }, + { + "epoch": 46.902255639097746, + "grad_norm": 0.023098763078451157, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31190 + }, + { + "epoch": 46.91729323308271, + "grad_norm": 0.032447148114442825, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 31200 + }, + { + "epoch": 46.932330827067666, + "grad_norm": 0.02725095860660076, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31210 + }, + { + "epoch": 46.94736842105263, + "grad_norm": 0.024231769144535065, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 31220 + }, + { + "epoch": 46.962406015037594, + "grad_norm": 0.026016365736722946, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31230 + }, + { + "epoch": 46.97744360902256, + "grad_norm": 0.017072904855012894, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31240 + }, + { + "epoch": 46.99248120300752, + "grad_norm": 0.03019643947482109, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31250 + }, + { + "epoch": 47.00751879699248, + "grad_norm": 0.032088860869407654, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 31260 + }, + { + "epoch": 47.02255639097744, + "grad_norm": 0.0365876741707325, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31270 + }, + { + "epoch": 47.037593984962406, + "grad_norm": 0.026992548257112503, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31280 + }, + { + "epoch": 47.05263157894737, + "grad_norm": 0.03195599839091301, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 31290 + }, + { + "epoch": 47.067669172932334, + "grad_norm": 0.02480645664036274, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 31300 + }, + { + "epoch": 47.08270676691729, + "grad_norm": 0.02834639884531498, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31310 + }, + { + "epoch": 47.097744360902254, + "grad_norm": 0.01962287537753582, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 31320 + }, + { + "epoch": 47.11278195488722, + "grad_norm": 0.013575027696788311, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31330 + }, + { + "epoch": 47.12781954887218, + "grad_norm": 0.03330725058913231, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31340 + }, + { + "epoch": 47.142857142857146, + "grad_norm": 0.020607665181159973, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31350 + }, + { + "epoch": 47.1578947368421, + "grad_norm": 0.023825455456972122, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31360 + }, + { + "epoch": 47.17293233082707, + "grad_norm": 0.018951276317238808, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31370 + }, + { + "epoch": 47.18796992481203, + "grad_norm": 0.03950987756252289, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31380 + }, + { + "epoch": 47.203007518796994, + "grad_norm": 0.01104127336293459, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31390 + }, + { + "epoch": 47.21804511278196, + "grad_norm": 0.025597169995307922, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31400 + }, + { + "epoch": 47.233082706766915, + "grad_norm": 0.017135456204414368, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31410 + }, + { + "epoch": 47.24812030075188, + "grad_norm": 0.021110041067004204, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31420 + }, + { + "epoch": 47.26315789473684, + "grad_norm": 0.020449955016374588, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31430 + }, + { + "epoch": 47.278195488721806, + "grad_norm": 0.021442551165819168, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31440 + }, + { + "epoch": 47.29323308270677, + "grad_norm": 0.02908310666680336, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31450 + }, + { + "epoch": 47.30827067669173, + "grad_norm": 0.023742174729704857, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31460 + }, + { + "epoch": 47.32330827067669, + "grad_norm": 0.0181658286601305, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31470 + }, + { + "epoch": 47.338345864661655, + "grad_norm": 0.03218434378504753, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31480 + }, + { + "epoch": 47.35338345864662, + "grad_norm": 0.025007104501128197, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 31490 + }, + { + "epoch": 47.36842105263158, + "grad_norm": 0.028918271884322166, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 31500 + }, + { + "epoch": 47.38345864661654, + "grad_norm": 0.030606761574745178, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31510 + }, + { + "epoch": 47.3984962406015, + "grad_norm": 0.02022801712155342, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31520 + }, + { + "epoch": 47.41353383458647, + "grad_norm": 0.03720887005329132, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31530 + }, + { + "epoch": 47.42857142857143, + "grad_norm": 0.03381654992699623, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31540 + }, + { + "epoch": 47.443609022556394, + "grad_norm": 0.023096594959497452, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31550 + }, + { + "epoch": 47.45864661654135, + "grad_norm": 0.029944781213998795, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31560 + }, + { + "epoch": 47.473684210526315, + "grad_norm": 0.01969045028090477, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31570 + }, + { + "epoch": 47.48872180451128, + "grad_norm": 0.04388013482093811, + "learning_rate": 0.0005, + "loss": 0.0023, + "step": 31580 + }, + { + "epoch": 47.50375939849624, + "grad_norm": 0.02591627836227417, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31590 + }, + { + "epoch": 47.5187969924812, + "grad_norm": 0.01688019372522831, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31600 + }, + { + "epoch": 47.53383458646616, + "grad_norm": 0.02898053079843521, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 31610 + }, + { + "epoch": 47.54887218045113, + "grad_norm": 0.02664492465555668, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31620 + }, + { + "epoch": 47.56390977443609, + "grad_norm": 0.014440646395087242, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 31630 + }, + { + "epoch": 47.578947368421055, + "grad_norm": 0.025000013411045074, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 31640 + }, + { + "epoch": 47.59398496240601, + "grad_norm": 0.03409901633858681, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31650 + }, + { + "epoch": 47.609022556390975, + "grad_norm": 0.027237460017204285, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31660 + }, + { + "epoch": 47.62406015037594, + "grad_norm": 0.025617195293307304, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31670 + }, + { + "epoch": 47.6390977443609, + "grad_norm": 0.018654363229870796, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31680 + }, + { + "epoch": 47.65413533834587, + "grad_norm": 0.030134858563542366, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31690 + }, + { + "epoch": 47.669172932330824, + "grad_norm": 0.01941750943660736, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31700 + }, + { + "epoch": 47.68421052631579, + "grad_norm": 0.017904341220855713, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31710 + }, + { + "epoch": 47.69924812030075, + "grad_norm": 0.02823181077837944, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31720 + }, + { + "epoch": 47.714285714285715, + "grad_norm": 0.031178584322333336, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31730 + }, + { + "epoch": 47.72932330827068, + "grad_norm": 0.03860757499933243, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31740 + }, + { + "epoch": 47.744360902255636, + "grad_norm": 0.034874770790338516, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31750 + }, + { + "epoch": 47.7593984962406, + "grad_norm": 0.02610880695283413, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 31760 + }, + { + "epoch": 47.774436090225564, + "grad_norm": 0.042836036533117294, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31770 + }, + { + "epoch": 47.78947368421053, + "grad_norm": 0.024372942745685577, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31780 + }, + { + "epoch": 47.80451127819549, + "grad_norm": 0.02181561291217804, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 31790 + }, + { + "epoch": 47.81954887218045, + "grad_norm": 0.029284097254276276, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 31800 + }, + { + "epoch": 47.83458646616541, + "grad_norm": 0.02451014332473278, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31810 + }, + { + "epoch": 47.849624060150376, + "grad_norm": 0.02723078243434429, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31820 + }, + { + "epoch": 47.86466165413534, + "grad_norm": 0.020348770543932915, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31830 + }, + { + "epoch": 47.8796992481203, + "grad_norm": 0.037076979875564575, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31840 + }, + { + "epoch": 47.89473684210526, + "grad_norm": 0.02210349589586258, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31850 + }, + { + "epoch": 47.909774436090224, + "grad_norm": 0.03658520057797432, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31860 + }, + { + "epoch": 47.92481203007519, + "grad_norm": 0.024562230333685875, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31870 + }, + { + "epoch": 47.93984962406015, + "grad_norm": 0.043723784387111664, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31880 + }, + { + "epoch": 47.954887218045116, + "grad_norm": 0.028547193855047226, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31890 + }, + { + "epoch": 47.96992481203007, + "grad_norm": 0.022432031109929085, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 31900 + }, + { + "epoch": 47.984962406015036, + "grad_norm": 0.03672812134027481, + "learning_rate": 0.0005, + "loss": 0.0021, + "step": 31910 + }, + { + "epoch": 48.0, + "grad_norm": 0.022083371877670288, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31920 + }, + { + "epoch": 48.015037593984964, + "grad_norm": 0.03273480385541916, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31930 + }, + { + "epoch": 48.03007518796993, + "grad_norm": 0.023940512910485268, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 31940 + }, + { + "epoch": 48.045112781954884, + "grad_norm": 0.040486034005880356, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31950 + }, + { + "epoch": 48.06015037593985, + "grad_norm": 0.014746240340173244, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31960 + }, + { + "epoch": 48.07518796992481, + "grad_norm": 0.030853888019919395, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 31970 + }, + { + "epoch": 48.090225563909776, + "grad_norm": 0.022572578862309456, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 31980 + }, + { + "epoch": 48.10526315789474, + "grad_norm": 0.020815961062908173, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 31990 + }, + { + "epoch": 48.1203007518797, + "grad_norm": 0.03605213016271591, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32000 + }, + { + "epoch": 48.1203007518797, + "eval_cer": 0.019365131610265506, + "eval_loss": 0.09814683347940445, + "eval_runtime": 162.4768, + "eval_samples_per_second": 98.642, + "eval_steps_per_second": 0.775, + "eval_wer": 0.0683812888797008, + "step": 32000 + }, + { + "epoch": 48.13533834586466, + "grad_norm": 0.03484274819493294, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32010 + }, + { + "epoch": 48.150375939849624, + "grad_norm": 0.023711401969194412, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32020 + }, + { + "epoch": 48.16541353383459, + "grad_norm": 0.019798826426267624, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32030 + }, + { + "epoch": 48.18045112781955, + "grad_norm": 0.017506971955299377, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32040 + }, + { + "epoch": 48.19548872180451, + "grad_norm": 0.017761750146746635, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32050 + }, + { + "epoch": 48.21052631578947, + "grad_norm": 0.016464564949274063, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32060 + }, + { + "epoch": 48.225563909774436, + "grad_norm": 0.02202432043850422, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32070 + }, + { + "epoch": 48.2406015037594, + "grad_norm": 0.030079545453190804, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32080 + }, + { + "epoch": 48.255639097744364, + "grad_norm": 0.038891155272722244, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32090 + }, + { + "epoch": 48.27067669172932, + "grad_norm": 0.027126679196953773, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32100 + }, + { + "epoch": 48.285714285714285, + "grad_norm": 0.025873566046357155, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32110 + }, + { + "epoch": 48.30075187969925, + "grad_norm": 0.029664838686585426, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32120 + }, + { + "epoch": 48.31578947368421, + "grad_norm": 0.02600649744272232, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32130 + }, + { + "epoch": 48.330827067669176, + "grad_norm": 0.024757621809840202, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 32140 + }, + { + "epoch": 48.34586466165413, + "grad_norm": 0.014252948574721813, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 32150 + }, + { + "epoch": 48.3609022556391, + "grad_norm": 0.017239531502127647, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32160 + }, + { + "epoch": 48.37593984962406, + "grad_norm": 0.015396323055028915, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32170 + }, + { + "epoch": 48.390977443609025, + "grad_norm": 0.022749189287424088, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32180 + }, + { + "epoch": 48.40601503759399, + "grad_norm": 0.033073458820581436, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32190 + }, + { + "epoch": 48.421052631578945, + "grad_norm": 0.01940765790641308, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32200 + }, + { + "epoch": 48.43609022556391, + "grad_norm": 0.018446555361151695, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32210 + }, + { + "epoch": 48.45112781954887, + "grad_norm": 0.026516687124967575, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32220 + }, + { + "epoch": 48.46616541353384, + "grad_norm": 0.03440384939312935, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32230 + }, + { + "epoch": 48.4812030075188, + "grad_norm": 0.03078754059970379, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32240 + }, + { + "epoch": 48.49624060150376, + "grad_norm": 0.03537391871213913, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 32250 + }, + { + "epoch": 48.51127819548872, + "grad_norm": 0.028442593291401863, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32260 + }, + { + "epoch": 48.526315789473685, + "grad_norm": 0.026463326066732407, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32270 + }, + { + "epoch": 48.54135338345865, + "grad_norm": 0.027073029428720474, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 32280 + }, + { + "epoch": 48.556390977443606, + "grad_norm": 0.023866435512900352, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32290 + }, + { + "epoch": 48.57142857142857, + "grad_norm": 0.029299238696694374, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32300 + }, + { + "epoch": 48.58646616541353, + "grad_norm": 0.027111824601888657, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32310 + }, + { + "epoch": 48.6015037593985, + "grad_norm": 0.008559158071875572, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32320 + }, + { + "epoch": 48.61654135338346, + "grad_norm": 0.01896803453564644, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32330 + }, + { + "epoch": 48.63157894736842, + "grad_norm": 0.03052617982029915, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32340 + }, + { + "epoch": 48.64661654135338, + "grad_norm": 0.03380419686436653, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32350 + }, + { + "epoch": 48.661654135338345, + "grad_norm": 0.02914763242006302, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 32360 + }, + { + "epoch": 48.67669172932331, + "grad_norm": 0.026141202077269554, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32370 + }, + { + "epoch": 48.69172932330827, + "grad_norm": 0.03368120267987251, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32380 + }, + { + "epoch": 48.70676691729323, + "grad_norm": 0.03151170536875725, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32390 + }, + { + "epoch": 48.721804511278194, + "grad_norm": 0.030847793444991112, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32400 + }, + { + "epoch": 48.73684210526316, + "grad_norm": 0.022047048434615135, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32410 + }, + { + "epoch": 48.75187969924812, + "grad_norm": 0.016951723024249077, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32420 + }, + { + "epoch": 48.766917293233085, + "grad_norm": 0.0469023771584034, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32430 + }, + { + "epoch": 48.78195488721804, + "grad_norm": 0.030872775241732597, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32440 + }, + { + "epoch": 48.796992481203006, + "grad_norm": 0.018543584272265434, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32450 + }, + { + "epoch": 48.81203007518797, + "grad_norm": 0.024176692590117455, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32460 + }, + { + "epoch": 48.82706766917293, + "grad_norm": 0.020511671900749207, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32470 + }, + { + "epoch": 48.8421052631579, + "grad_norm": 0.019052179530262947, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32480 + }, + { + "epoch": 48.857142857142854, + "grad_norm": 0.028492294251918793, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32490 + }, + { + "epoch": 48.87218045112782, + "grad_norm": 0.020564468577504158, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32500 + }, + { + "epoch": 48.88721804511278, + "grad_norm": 0.03670753166079521, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32510 + }, + { + "epoch": 48.902255639097746, + "grad_norm": 0.0216042622923851, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32520 + }, + { + "epoch": 48.91729323308271, + "grad_norm": 0.019789384678006172, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 32530 + }, + { + "epoch": 48.932330827067666, + "grad_norm": 0.025021541863679886, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 32540 + }, + { + "epoch": 48.94736842105263, + "grad_norm": 0.041851602494716644, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 32550 + }, + { + "epoch": 48.962406015037594, + "grad_norm": 0.023716002702713013, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32560 + }, + { + "epoch": 48.97744360902256, + "grad_norm": 0.027679838240146637, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32570 + }, + { + "epoch": 48.99248120300752, + "grad_norm": 0.02079549804329872, + "learning_rate": 0.0005, + "loss": 0.002, + "step": 32580 + }, + { + "epoch": 49.00751879699248, + "grad_norm": 0.03486000373959541, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32590 + }, + { + "epoch": 49.02255639097744, + "grad_norm": 0.02394845522940159, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 32600 + }, + { + "epoch": 49.037593984962406, + "grad_norm": 0.030554912984371185, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32610 + }, + { + "epoch": 49.05263157894737, + "grad_norm": 0.01994483731687069, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 32620 + }, + { + "epoch": 49.067669172932334, + "grad_norm": 0.027919717133045197, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32630 + }, + { + "epoch": 49.08270676691729, + "grad_norm": 0.021079905331134796, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32640 + }, + { + "epoch": 49.097744360902254, + "grad_norm": 0.026793304830789566, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32650 + }, + { + "epoch": 49.11278195488722, + "grad_norm": 0.017919959500432014, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32660 + }, + { + "epoch": 49.12781954887218, + "grad_norm": 0.017721813172101974, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32670 + }, + { + "epoch": 49.142857142857146, + "grad_norm": 0.043451085686683655, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 32680 + }, + { + "epoch": 49.1578947368421, + "grad_norm": 0.02092374488711357, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32690 + }, + { + "epoch": 49.17293233082707, + "grad_norm": 0.046850405633449554, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 32700 + }, + { + "epoch": 49.18796992481203, + "grad_norm": 0.026273690164089203, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32710 + }, + { + "epoch": 49.203007518796994, + "grad_norm": 0.015484370291233063, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 32720 + }, + { + "epoch": 49.21804511278196, + "grad_norm": 0.015919826924800873, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32730 + }, + { + "epoch": 49.233082706766915, + "grad_norm": 0.027749303728342056, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32740 + }, + { + "epoch": 49.24812030075188, + "grad_norm": 0.022383715957403183, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 32750 + }, + { + "epoch": 49.26315789473684, + "grad_norm": 0.02893335558474064, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32760 + }, + { + "epoch": 49.278195488721806, + "grad_norm": 0.033238474279642105, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32770 + }, + { + "epoch": 49.29323308270677, + "grad_norm": 0.016897819936275482, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32780 + }, + { + "epoch": 49.30827067669173, + "grad_norm": 0.049246326088905334, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32790 + }, + { + "epoch": 49.32330827067669, + "grad_norm": 0.02789200097322464, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 32800 + }, + { + "epoch": 49.338345864661655, + "grad_norm": 0.012685381807386875, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32810 + }, + { + "epoch": 49.35338345864662, + "grad_norm": 0.02751254476606846, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32820 + }, + { + "epoch": 49.36842105263158, + "grad_norm": 0.023942092433571815, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32830 + }, + { + "epoch": 49.38345864661654, + "grad_norm": 0.02301846072077751, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32840 + }, + { + "epoch": 49.3984962406015, + "grad_norm": 0.03137141093611717, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32850 + }, + { + "epoch": 49.41353383458647, + "grad_norm": 0.01922498270869255, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 32860 + }, + { + "epoch": 49.42857142857143, + "grad_norm": 0.04007174074649811, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32870 + }, + { + "epoch": 49.443609022556394, + "grad_norm": 0.03627806156873703, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32880 + }, + { + "epoch": 49.45864661654135, + "grad_norm": 0.02141883224248886, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32890 + }, + { + "epoch": 49.473684210526315, + "grad_norm": 0.02805938571691513, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 32900 + }, + { + "epoch": 49.48872180451128, + "grad_norm": 0.014750118367373943, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32910 + }, + { + "epoch": 49.50375939849624, + "grad_norm": 0.020815087482333183, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 32920 + }, + { + "epoch": 49.5187969924812, + "grad_norm": 0.022273480892181396, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32930 + }, + { + "epoch": 49.53383458646616, + "grad_norm": 0.029769672080874443, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32940 + }, + { + "epoch": 49.54887218045113, + "grad_norm": 0.01716325432062149, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 32950 + }, + { + "epoch": 49.56390977443609, + "grad_norm": 0.031024549156427383, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 32960 + }, + { + "epoch": 49.578947368421055, + "grad_norm": 0.03407551720738411, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 32970 + }, + { + "epoch": 49.59398496240601, + "grad_norm": 0.02655191719532013, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32980 + }, + { + "epoch": 49.609022556390975, + "grad_norm": 0.01913565769791603, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 32990 + }, + { + "epoch": 49.62406015037594, + "grad_norm": 0.02799713984131813, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33000 + }, + { + "epoch": 49.62406015037594, + "eval_cer": 0.01911419279989228, + "eval_loss": 0.09544375538825989, + "eval_runtime": 161.6046, + "eval_samples_per_second": 99.174, + "eval_steps_per_second": 0.78, + "eval_wer": 0.0679490311789359, + "step": 33000 + }, + { + "epoch": 49.6390977443609, + "grad_norm": 0.02007681131362915, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33010 + }, + { + "epoch": 49.65413533834587, + "grad_norm": 0.024770651012659073, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33020 + }, + { + "epoch": 49.669172932330824, + "grad_norm": 0.025589197874069214, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 33030 + }, + { + "epoch": 49.68421052631579, + "grad_norm": 0.03725748509168625, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33040 + }, + { + "epoch": 49.69924812030075, + "grad_norm": 0.03386741504073143, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33050 + }, + { + "epoch": 49.714285714285715, + "grad_norm": 0.023790443316102028, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33060 + }, + { + "epoch": 49.72932330827068, + "grad_norm": 0.0374760664999485, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33070 + }, + { + "epoch": 49.744360902255636, + "grad_norm": 0.01530708558857441, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33080 + }, + { + "epoch": 49.7593984962406, + "grad_norm": 0.022537345066666603, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33090 + }, + { + "epoch": 49.774436090225564, + "grad_norm": 0.02101951092481613, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33100 + }, + { + "epoch": 49.78947368421053, + "grad_norm": 0.02951240725815296, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33110 + }, + { + "epoch": 49.80451127819549, + "grad_norm": 0.023769032210111618, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33120 + }, + { + "epoch": 49.81954887218045, + "grad_norm": 0.027803754433989525, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33130 + }, + { + "epoch": 49.83458646616541, + "grad_norm": 0.02038266696035862, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33140 + }, + { + "epoch": 49.849624060150376, + "grad_norm": 0.024007432162761688, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33150 + }, + { + "epoch": 49.86466165413534, + "grad_norm": 0.021218260750174522, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33160 + }, + { + "epoch": 49.8796992481203, + "grad_norm": 0.02512981742620468, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33170 + }, + { + "epoch": 49.89473684210526, + "grad_norm": 0.017900394275784492, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33180 + }, + { + "epoch": 49.909774436090224, + "grad_norm": 0.022142210975289345, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33190 + }, + { + "epoch": 49.92481203007519, + "grad_norm": 0.016763297840952873, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 33200 + }, + { + "epoch": 49.93984962406015, + "grad_norm": 0.01720983348786831, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33210 + }, + { + "epoch": 49.954887218045116, + "grad_norm": 0.031225530430674553, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 33220 + }, + { + "epoch": 49.96992481203007, + "grad_norm": 0.016232699155807495, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33230 + }, + { + "epoch": 49.984962406015036, + "grad_norm": 0.026164473965764046, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 33240 + }, + { + "epoch": 50.0, + "grad_norm": 0.04493225738406181, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33250 + }, + { + "epoch": 50.015037593984964, + "grad_norm": 0.020769275724887848, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33260 + }, + { + "epoch": 50.03007518796993, + "grad_norm": 0.009784534573554993, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33270 + }, + { + "epoch": 50.045112781954884, + "grad_norm": 0.038556769490242004, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33280 + }, + { + "epoch": 50.06015037593985, + "grad_norm": 0.01689024828374386, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33290 + }, + { + "epoch": 50.07518796992481, + "grad_norm": 0.0235456433147192, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33300 + }, + { + "epoch": 50.090225563909776, + "grad_norm": 0.013643356040120125, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33310 + }, + { + "epoch": 50.10526315789474, + "grad_norm": 0.030662594363093376, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33320 + }, + { + "epoch": 50.1203007518797, + "grad_norm": 0.032547783106565475, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 33330 + }, + { + "epoch": 50.13533834586466, + "grad_norm": 0.029664169996976852, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33340 + }, + { + "epoch": 50.150375939849624, + "grad_norm": 0.030851969495415688, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33350 + }, + { + "epoch": 50.16541353383459, + "grad_norm": 0.02700444683432579, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33360 + }, + { + "epoch": 50.18045112781955, + "grad_norm": 0.027207769453525543, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33370 + }, + { + "epoch": 50.19548872180451, + "grad_norm": 0.018520159646868706, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 33380 + }, + { + "epoch": 50.21052631578947, + "grad_norm": 0.018409132957458496, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33390 + }, + { + "epoch": 50.225563909774436, + "grad_norm": 0.03714638575911522, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33400 + }, + { + "epoch": 50.2406015037594, + "grad_norm": 0.023851973935961723, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 33410 + }, + { + "epoch": 50.255639097744364, + "grad_norm": 0.0141056589782238, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33420 + }, + { + "epoch": 50.27067669172932, + "grad_norm": 0.02931864932179451, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33430 + }, + { + "epoch": 50.285714285714285, + "grad_norm": 0.015516453422605991, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33440 + }, + { + "epoch": 50.30075187969925, + "grad_norm": 0.02239651419222355, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33450 + }, + { + "epoch": 50.31578947368421, + "grad_norm": 0.014235914684832096, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 33460 + }, + { + "epoch": 50.330827067669176, + "grad_norm": 0.02175496704876423, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33470 + }, + { + "epoch": 50.34586466165413, + "grad_norm": 0.01744878478348255, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33480 + }, + { + "epoch": 50.3609022556391, + "grad_norm": 0.03548421338200569, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33490 + }, + { + "epoch": 50.37593984962406, + "grad_norm": 0.029608802869915962, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 33500 + }, + { + "epoch": 50.390977443609025, + "grad_norm": 0.021057721227407455, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33510 + }, + { + "epoch": 50.40601503759399, + "grad_norm": 0.022127574309706688, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33520 + }, + { + "epoch": 50.421052631578945, + "grad_norm": 0.027659917250275612, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 33530 + }, + { + "epoch": 50.43609022556391, + "grad_norm": 0.018252188339829445, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33540 + }, + { + "epoch": 50.45112781954887, + "grad_norm": 0.028814923018217087, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33550 + }, + { + "epoch": 50.46616541353384, + "grad_norm": 0.028050053864717484, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33560 + }, + { + "epoch": 50.4812030075188, + "grad_norm": 0.02291923388838768, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33570 + }, + { + "epoch": 50.49624060150376, + "grad_norm": 0.026604672893881798, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33580 + }, + { + "epoch": 50.51127819548872, + "grad_norm": 0.02081146650016308, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 33590 + }, + { + "epoch": 50.526315789473685, + "grad_norm": 0.021414954215288162, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33600 + }, + { + "epoch": 50.54135338345865, + "grad_norm": 0.02391757071018219, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33610 + }, + { + "epoch": 50.556390977443606, + "grad_norm": 0.021122314035892487, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33620 + }, + { + "epoch": 50.57142857142857, + "grad_norm": 0.043358538299798965, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33630 + }, + { + "epoch": 50.58646616541353, + "grad_norm": 0.022805282846093178, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33640 + }, + { + "epoch": 50.6015037593985, + "grad_norm": 0.017654482275247574, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33650 + }, + { + "epoch": 50.61654135338346, + "grad_norm": 0.022442620247602463, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33660 + }, + { + "epoch": 50.63157894736842, + "grad_norm": 0.015780918300151825, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33670 + }, + { + "epoch": 50.64661654135338, + "grad_norm": 0.025658588856458664, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 33680 + }, + { + "epoch": 50.661654135338345, + "grad_norm": 0.032824546098709106, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33690 + }, + { + "epoch": 50.67669172932331, + "grad_norm": 0.02054538019001484, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33700 + }, + { + "epoch": 50.69172932330827, + "grad_norm": 0.02426365576684475, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33710 + }, + { + "epoch": 50.70676691729323, + "grad_norm": 0.019750064238905907, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33720 + }, + { + "epoch": 50.721804511278194, + "grad_norm": 0.019385552033782005, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33730 + }, + { + "epoch": 50.73684210526316, + "grad_norm": 0.021297238767147064, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33740 + }, + { + "epoch": 50.75187969924812, + "grad_norm": 0.03146389499306679, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33750 + }, + { + "epoch": 50.766917293233085, + "grad_norm": 0.044092606753110886, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 33760 + }, + { + "epoch": 50.78195488721804, + "grad_norm": 0.0318281389772892, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33770 + }, + { + "epoch": 50.796992481203006, + "grad_norm": 0.02030392549932003, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33780 + }, + { + "epoch": 50.81203007518797, + "grad_norm": 0.037234265357255936, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 33790 + }, + { + "epoch": 50.82706766917293, + "grad_norm": 0.02704983949661255, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33800 + }, + { + "epoch": 50.8421052631579, + "grad_norm": 0.026450637727975845, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33810 + }, + { + "epoch": 50.857142857142854, + "grad_norm": 0.026690615341067314, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33820 + }, + { + "epoch": 50.87218045112782, + "grad_norm": 0.025579355657100677, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33830 + }, + { + "epoch": 50.88721804511278, + "grad_norm": 0.031650979071855545, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 33840 + }, + { + "epoch": 50.902255639097746, + "grad_norm": 0.026199985295534134, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 33850 + }, + { + "epoch": 50.91729323308271, + "grad_norm": 0.025939514860510826, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33860 + }, + { + "epoch": 50.932330827067666, + "grad_norm": 0.024666165933012962, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 33870 + }, + { + "epoch": 50.94736842105263, + "grad_norm": 0.021048100665211678, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33880 + }, + { + "epoch": 50.962406015037594, + "grad_norm": 0.02089071087539196, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33890 + }, + { + "epoch": 50.97744360902256, + "grad_norm": 0.019832544028759003, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33900 + }, + { + "epoch": 50.99248120300752, + "grad_norm": 0.02231118269264698, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 33910 + }, + { + "epoch": 51.00751879699248, + "grad_norm": 0.022497259080410004, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33920 + }, + { + "epoch": 51.02255639097744, + "grad_norm": 0.024080676957964897, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33930 + }, + { + "epoch": 51.037593984962406, + "grad_norm": 0.01972859725356102, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 33940 + }, + { + "epoch": 51.05263157894737, + "grad_norm": 0.019077617675065994, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 33950 + }, + { + "epoch": 51.067669172932334, + "grad_norm": 0.030107924714684486, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 33960 + }, + { + "epoch": 51.08270676691729, + "grad_norm": 0.02160046249628067, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 33970 + }, + { + "epoch": 51.097744360902254, + "grad_norm": 0.016069544479250908, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 33980 + }, + { + "epoch": 51.11278195488722, + "grad_norm": 0.03132305294275284, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 33990 + }, + { + "epoch": 51.12781954887218, + "grad_norm": 0.02570117823779583, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34000 + }, + { + "epoch": 51.12781954887218, + "eval_cer": 0.018937379545970766, + "eval_loss": 0.09406783431768417, + "eval_runtime": 158.0488, + "eval_samples_per_second": 101.405, + "eval_steps_per_second": 0.797, + "eval_wer": 0.06867729143565938, + "step": 34000 + }, + { + "epoch": 51.142857142857146, + "grad_norm": 0.016079319640994072, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34010 + }, + { + "epoch": 51.1578947368421, + "grad_norm": 0.025908542796969414, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34020 + }, + { + "epoch": 51.17293233082707, + "grad_norm": 0.02018142305314541, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34030 + }, + { + "epoch": 51.18796992481203, + "grad_norm": 0.016095496714115143, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34040 + }, + { + "epoch": 51.203007518796994, + "grad_norm": 0.016679896041750908, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34050 + }, + { + "epoch": 51.21804511278196, + "grad_norm": 0.032604388892650604, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34060 + }, + { + "epoch": 51.233082706766915, + "grad_norm": 0.021062975749373436, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34070 + }, + { + "epoch": 51.24812030075188, + "grad_norm": 0.02118678018450737, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34080 + }, + { + "epoch": 51.26315789473684, + "grad_norm": 0.038812920451164246, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34090 + }, + { + "epoch": 51.278195488721806, + "grad_norm": 0.01823197677731514, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34100 + }, + { + "epoch": 51.29323308270677, + "grad_norm": 0.029164424166083336, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34110 + }, + { + "epoch": 51.30827067669173, + "grad_norm": 0.034905001521110535, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 34120 + }, + { + "epoch": 51.32330827067669, + "grad_norm": 0.022618917748332024, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34130 + }, + { + "epoch": 51.338345864661655, + "grad_norm": 0.033336859196424484, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34140 + }, + { + "epoch": 51.35338345864662, + "grad_norm": 0.02516194060444832, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34150 + }, + { + "epoch": 51.36842105263158, + "grad_norm": 0.03591986373066902, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34160 + }, + { + "epoch": 51.38345864661654, + "grad_norm": 0.02537713013589382, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34170 + }, + { + "epoch": 51.3984962406015, + "grad_norm": 0.02011830545961857, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34180 + }, + { + "epoch": 51.41353383458647, + "grad_norm": 0.03309379145503044, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34190 + }, + { + "epoch": 51.42857142857143, + "grad_norm": 0.016980476677417755, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34200 + }, + { + "epoch": 51.443609022556394, + "grad_norm": 0.020793598145246506, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34210 + }, + { + "epoch": 51.45864661654135, + "grad_norm": 0.01948828622698784, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34220 + }, + { + "epoch": 51.473684210526315, + "grad_norm": 0.024978838860988617, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34230 + }, + { + "epoch": 51.48872180451128, + "grad_norm": 0.02950778789818287, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34240 + }, + { + "epoch": 51.50375939849624, + "grad_norm": 0.01592310518026352, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34250 + }, + { + "epoch": 51.5187969924812, + "grad_norm": 0.02595733106136322, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34260 + }, + { + "epoch": 51.53383458646616, + "grad_norm": 0.021075541153550148, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34270 + }, + { + "epoch": 51.54887218045113, + "grad_norm": 0.02784183993935585, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34280 + }, + { + "epoch": 51.56390977443609, + "grad_norm": 0.01684662327170372, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34290 + }, + { + "epoch": 51.578947368421055, + "grad_norm": 0.021374458447098732, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34300 + }, + { + "epoch": 51.59398496240601, + "grad_norm": 0.01987413503229618, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34310 + }, + { + "epoch": 51.609022556390975, + "grad_norm": 0.02309625782072544, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34320 + }, + { + "epoch": 51.62406015037594, + "grad_norm": 0.022111788392066956, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 34330 + }, + { + "epoch": 51.6390977443609, + "grad_norm": 0.014152990654110909, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 34340 + }, + { + "epoch": 51.65413533834587, + "grad_norm": 0.014807388186454773, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34350 + }, + { + "epoch": 51.669172932330824, + "grad_norm": 0.010940113104879856, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34360 + }, + { + "epoch": 51.68421052631579, + "grad_norm": 0.021173721179366112, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34370 + }, + { + "epoch": 51.69924812030075, + "grad_norm": 0.02962423302233219, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34380 + }, + { + "epoch": 51.714285714285715, + "grad_norm": 0.02886929176747799, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34390 + }, + { + "epoch": 51.72932330827068, + "grad_norm": 0.017438694834709167, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34400 + }, + { + "epoch": 51.744360902255636, + "grad_norm": 0.026239853352308273, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34410 + }, + { + "epoch": 51.7593984962406, + "grad_norm": 0.02102925442159176, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34420 + }, + { + "epoch": 51.774436090225564, + "grad_norm": 0.03565576300024986, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34430 + }, + { + "epoch": 51.78947368421053, + "grad_norm": 0.01991630159318447, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34440 + }, + { + "epoch": 51.80451127819549, + "grad_norm": 0.020549364387989044, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34450 + }, + { + "epoch": 51.81954887218045, + "grad_norm": 0.021338626742362976, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34460 + }, + { + "epoch": 51.83458646616541, + "grad_norm": 0.019852371886372566, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34470 + }, + { + "epoch": 51.849624060150376, + "grad_norm": 0.015359841287136078, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34480 + }, + { + "epoch": 51.86466165413534, + "grad_norm": 0.029628610238432884, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34490 + }, + { + "epoch": 51.8796992481203, + "grad_norm": 0.027623649686574936, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34500 + }, + { + "epoch": 51.89473684210526, + "grad_norm": 0.03379254788160324, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34510 + }, + { + "epoch": 51.909774436090224, + "grad_norm": 0.036563072353601456, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34520 + }, + { + "epoch": 51.92481203007519, + "grad_norm": 0.01687576062977314, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34530 + }, + { + "epoch": 51.93984962406015, + "grad_norm": 0.019457759335637093, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34540 + }, + { + "epoch": 51.954887218045116, + "grad_norm": 0.03398590534925461, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34550 + }, + { + "epoch": 51.96992481203007, + "grad_norm": 0.021139761433005333, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34560 + }, + { + "epoch": 51.984962406015036, + "grad_norm": 0.04651300236582756, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34570 + }, + { + "epoch": 52.0, + "grad_norm": 0.024219822138547897, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34580 + }, + { + "epoch": 52.015037593984964, + "grad_norm": 0.023489603772759438, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34590 + }, + { + "epoch": 52.03007518796993, + "grad_norm": 0.020888742059469223, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 34600 + }, + { + "epoch": 52.045112781954884, + "grad_norm": 0.030688663944602013, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34610 + }, + { + "epoch": 52.06015037593985, + "grad_norm": 0.01604343391954899, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34620 + }, + { + "epoch": 52.07518796992481, + "grad_norm": 0.04560728743672371, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34630 + }, + { + "epoch": 52.090225563909776, + "grad_norm": 0.015305062755942345, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34640 + }, + { + "epoch": 52.10526315789474, + "grad_norm": 0.02733980119228363, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34650 + }, + { + "epoch": 52.1203007518797, + "grad_norm": 0.018215160816907883, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 34660 + }, + { + "epoch": 52.13533834586466, + "grad_norm": 0.0145012978464365, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34670 + }, + { + "epoch": 52.150375939849624, + "grad_norm": 0.03274387866258621, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34680 + }, + { + "epoch": 52.16541353383459, + "grad_norm": 0.026213057339191437, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34690 + }, + { + "epoch": 52.18045112781955, + "grad_norm": 0.01359515730291605, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34700 + }, + { + "epoch": 52.19548872180451, + "grad_norm": 0.021538972854614258, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34710 + }, + { + "epoch": 52.21052631578947, + "grad_norm": 0.017255106940865517, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34720 + }, + { + "epoch": 52.225563909774436, + "grad_norm": 0.02000657096505165, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34730 + }, + { + "epoch": 52.2406015037594, + "grad_norm": 0.02495522052049637, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34740 + }, + { + "epoch": 52.255639097744364, + "grad_norm": 0.011161725968122482, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34750 + }, + { + "epoch": 52.27067669172932, + "grad_norm": 0.02745555341243744, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34760 + }, + { + "epoch": 52.285714285714285, + "grad_norm": 0.018100611865520477, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34770 + }, + { + "epoch": 52.30075187969925, + "grad_norm": 0.022142384201288223, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34780 + }, + { + "epoch": 52.31578947368421, + "grad_norm": 0.038946229964494705, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34790 + }, + { + "epoch": 52.330827067669176, + "grad_norm": 0.02243795432150364, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 34800 + }, + { + "epoch": 52.34586466165413, + "grad_norm": 0.015783054754137993, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34810 + }, + { + "epoch": 52.3609022556391, + "grad_norm": 0.03521750494837761, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34820 + }, + { + "epoch": 52.37593984962406, + "grad_norm": 0.028525812551379204, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34830 + }, + { + "epoch": 52.390977443609025, + "grad_norm": 0.025964660570025444, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34840 + }, + { + "epoch": 52.40601503759399, + "grad_norm": 0.026434658095240593, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34850 + }, + { + "epoch": 52.421052631578945, + "grad_norm": 0.031102558597922325, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 34860 + }, + { + "epoch": 52.43609022556391, + "grad_norm": 0.026999972760677338, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34870 + }, + { + "epoch": 52.45112781954887, + "grad_norm": 0.02129795029759407, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34880 + }, + { + "epoch": 52.46616541353384, + "grad_norm": 0.03257288411259651, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34890 + }, + { + "epoch": 52.4812030075188, + "grad_norm": 0.03901764377951622, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34900 + }, + { + "epoch": 52.49624060150376, + "grad_norm": 0.0321793369948864, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 34910 + }, + { + "epoch": 52.51127819548872, + "grad_norm": 0.021580923348665237, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34920 + }, + { + "epoch": 52.526315789473685, + "grad_norm": 0.025633666664361954, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34930 + }, + { + "epoch": 52.54135338345865, + "grad_norm": 0.020235726609826088, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34940 + }, + { + "epoch": 52.556390977443606, + "grad_norm": 0.035812072455883026, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 34950 + }, + { + "epoch": 52.57142857142857, + "grad_norm": 0.022313429042696953, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 34960 + }, + { + "epoch": 52.58646616541353, + "grad_norm": 0.022014133632183075, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 34970 + }, + { + "epoch": 52.6015037593985, + "grad_norm": 0.032092828303575516, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 34980 + }, + { + "epoch": 52.61654135338346, + "grad_norm": 0.018840495496988297, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 34990 + }, + { + "epoch": 52.63157894736842, + "grad_norm": 0.01085710246115923, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35000 + }, + { + "epoch": 52.63157894736842, + "eval_cer": 0.01850622722679292, + "eval_loss": 0.09714452177286148, + "eval_runtime": 160.8173, + "eval_samples_per_second": 99.66, + "eval_steps_per_second": 0.783, + "eval_wer": 0.06823563682835611, + "step": 35000 + }, + { + "epoch": 52.64661654135338, + "grad_norm": 0.029773874208331108, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 35010 + }, + { + "epoch": 52.661654135338345, + "grad_norm": 0.023813791573047638, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35020 + }, + { + "epoch": 52.67669172932331, + "grad_norm": 0.02223018929362297, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35030 + }, + { + "epoch": 52.69172932330827, + "grad_norm": 0.016451871022582054, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35040 + }, + { + "epoch": 52.70676691729323, + "grad_norm": 0.03455843776464462, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35050 + }, + { + "epoch": 52.721804511278194, + "grad_norm": 0.0269955825060606, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35060 + }, + { + "epoch": 52.73684210526316, + "grad_norm": 0.03141428902745247, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35070 + }, + { + "epoch": 52.75187969924812, + "grad_norm": 0.022005867213010788, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35080 + }, + { + "epoch": 52.766917293233085, + "grad_norm": 0.03449671342968941, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35090 + }, + { + "epoch": 52.78195488721804, + "grad_norm": 0.03554965928196907, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35100 + }, + { + "epoch": 52.796992481203006, + "grad_norm": 0.018003834411501884, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35110 + }, + { + "epoch": 52.81203007518797, + "grad_norm": 0.020208600908517838, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35120 + }, + { + "epoch": 52.82706766917293, + "grad_norm": 0.01983526349067688, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35130 + }, + { + "epoch": 52.8421052631579, + "grad_norm": 0.03591633215546608, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 35140 + }, + { + "epoch": 52.857142857142854, + "grad_norm": 0.0240222979336977, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35150 + }, + { + "epoch": 52.87218045112782, + "grad_norm": 0.02874290756881237, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 35160 + }, + { + "epoch": 52.88721804511278, + "grad_norm": 0.020319845527410507, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 35170 + }, + { + "epoch": 52.902255639097746, + "grad_norm": 0.013919001445174217, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 35180 + }, + { + "epoch": 52.91729323308271, + "grad_norm": 0.022105760872364044, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35190 + }, + { + "epoch": 52.932330827067666, + "grad_norm": 0.02879832126200199, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35200 + }, + { + "epoch": 52.94736842105263, + "grad_norm": 0.0327581986784935, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35210 + }, + { + "epoch": 52.962406015037594, + "grad_norm": 0.04804421216249466, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35220 + }, + { + "epoch": 52.97744360902256, + "grad_norm": 0.015865134075284004, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35230 + }, + { + "epoch": 52.99248120300752, + "grad_norm": 0.029054423794150352, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35240 + }, + { + "epoch": 53.00751879699248, + "grad_norm": 0.02804388478398323, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 35250 + }, + { + "epoch": 53.02255639097744, + "grad_norm": 0.019160345196723938, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35260 + }, + { + "epoch": 53.037593984962406, + "grad_norm": 0.023717382922768593, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35270 + }, + { + "epoch": 53.05263157894737, + "grad_norm": 0.02732754312455654, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35280 + }, + { + "epoch": 53.067669172932334, + "grad_norm": 0.016642188653349876, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 35290 + }, + { + "epoch": 53.08270676691729, + "grad_norm": 0.027704434469342232, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35300 + }, + { + "epoch": 53.097744360902254, + "grad_norm": 0.02447352558374405, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35310 + }, + { + "epoch": 53.11278195488722, + "grad_norm": 0.027852097526192665, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35320 + }, + { + "epoch": 53.12781954887218, + "grad_norm": 0.029805408790707588, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35330 + }, + { + "epoch": 53.142857142857146, + "grad_norm": 0.022506574168801308, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35340 + }, + { + "epoch": 53.1578947368421, + "grad_norm": 0.03144741430878639, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35350 + }, + { + "epoch": 53.17293233082707, + "grad_norm": 0.01587393693625927, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35360 + }, + { + "epoch": 53.18796992481203, + "grad_norm": 0.0190495066344738, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35370 + }, + { + "epoch": 53.203007518796994, + "grad_norm": 0.020844964310526848, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35380 + }, + { + "epoch": 53.21804511278196, + "grad_norm": 0.021898053586483, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35390 + }, + { + "epoch": 53.233082706766915, + "grad_norm": 0.023180026561021805, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 35400 + }, + { + "epoch": 53.24812030075188, + "grad_norm": 0.024993926286697388, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35410 + }, + { + "epoch": 53.26315789473684, + "grad_norm": 0.020093608647584915, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35420 + }, + { + "epoch": 53.278195488721806, + "grad_norm": 0.020982353016734123, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35430 + }, + { + "epoch": 53.29323308270677, + "grad_norm": 0.0285146813839674, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35440 + }, + { + "epoch": 53.30827067669173, + "grad_norm": 0.031351540237665176, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35450 + }, + { + "epoch": 53.32330827067669, + "grad_norm": 0.01689954288303852, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35460 + }, + { + "epoch": 53.338345864661655, + "grad_norm": 0.012142537161707878, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 35470 + }, + { + "epoch": 53.35338345864662, + "grad_norm": 0.025104453787207603, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35480 + }, + { + "epoch": 53.36842105263158, + "grad_norm": 0.0306001678109169, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35490 + }, + { + "epoch": 53.38345864661654, + "grad_norm": 0.03842427581548691, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35500 + }, + { + "epoch": 53.3984962406015, + "grad_norm": 0.026987554505467415, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35510 + }, + { + "epoch": 53.41353383458647, + "grad_norm": 0.029108582064509392, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35520 + }, + { + "epoch": 53.42857142857143, + "grad_norm": 0.026543794199824333, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35530 + }, + { + "epoch": 53.443609022556394, + "grad_norm": 0.0192901398986578, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35540 + }, + { + "epoch": 53.45864661654135, + "grad_norm": 0.02008858136832714, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35550 + }, + { + "epoch": 53.473684210526315, + "grad_norm": 0.033383749425411224, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 35560 + }, + { + "epoch": 53.48872180451128, + "grad_norm": 0.023570675402879715, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35570 + }, + { + "epoch": 53.50375939849624, + "grad_norm": 0.033952970057725906, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35580 + }, + { + "epoch": 53.5187969924812, + "grad_norm": 0.011865677312016487, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35590 + }, + { + "epoch": 53.53383458646616, + "grad_norm": 0.020529886707663536, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35600 + }, + { + "epoch": 53.54887218045113, + "grad_norm": 0.03204064816236496, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35610 + }, + { + "epoch": 53.56390977443609, + "grad_norm": 0.01883283071219921, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 35620 + }, + { + "epoch": 53.578947368421055, + "grad_norm": 0.017379991710186005, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35630 + }, + { + "epoch": 53.59398496240601, + "grad_norm": 0.025450879707932472, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35640 + }, + { + "epoch": 53.609022556390975, + "grad_norm": 0.022188102826476097, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35650 + }, + { + "epoch": 53.62406015037594, + "grad_norm": 0.027863452211022377, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35660 + }, + { + "epoch": 53.6390977443609, + "grad_norm": 0.026337917894124985, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35670 + }, + { + "epoch": 53.65413533834587, + "grad_norm": 0.017244908958673477, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 35680 + }, + { + "epoch": 53.669172932330824, + "grad_norm": 0.043919071555137634, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35690 + }, + { + "epoch": 53.68421052631579, + "grad_norm": 0.029560135677456856, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35700 + }, + { + "epoch": 53.69924812030075, + "grad_norm": 0.03068513609468937, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35710 + }, + { + "epoch": 53.714285714285715, + "grad_norm": 0.03232008218765259, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35720 + }, + { + "epoch": 53.72932330827068, + "grad_norm": 0.022236887365579605, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35730 + }, + { + "epoch": 53.744360902255636, + "grad_norm": 0.029042409732937813, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35740 + }, + { + "epoch": 53.7593984962406, + "grad_norm": 0.029713962227106094, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35750 + }, + { + "epoch": 53.774436090225564, + "grad_norm": 0.015098396688699722, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35760 + }, + { + "epoch": 53.78947368421053, + "grad_norm": 0.018189474940299988, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35770 + }, + { + "epoch": 53.80451127819549, + "grad_norm": 0.016704337671399117, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 35780 + }, + { + "epoch": 53.81954887218045, + "grad_norm": 0.026412488892674446, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35790 + }, + { + "epoch": 53.83458646616541, + "grad_norm": 0.021768754348158836, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35800 + }, + { + "epoch": 53.849624060150376, + "grad_norm": 0.031175939366221428, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35810 + }, + { + "epoch": 53.86466165413534, + "grad_norm": 0.029222097247838974, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35820 + }, + { + "epoch": 53.8796992481203, + "grad_norm": 0.023655666038393974, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35830 + }, + { + "epoch": 53.89473684210526, + "grad_norm": 0.022948008030653, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35840 + }, + { + "epoch": 53.909774436090224, + "grad_norm": 0.024354655295610428, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35850 + }, + { + "epoch": 53.92481203007519, + "grad_norm": 0.016911059617996216, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35860 + }, + { + "epoch": 53.93984962406015, + "grad_norm": 0.023537425324320793, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 35870 + }, + { + "epoch": 53.954887218045116, + "grad_norm": 0.04137105867266655, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35880 + }, + { + "epoch": 53.96992481203007, + "grad_norm": 0.023718271404504776, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35890 + }, + { + "epoch": 53.984962406015036, + "grad_norm": 0.029552804306149483, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 35900 + }, + { + "epoch": 54.0, + "grad_norm": 0.020634911954402924, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 35910 + }, + { + "epoch": 54.015037593984964, + "grad_norm": 0.02160995453596115, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35920 + }, + { + "epoch": 54.03007518796993, + "grad_norm": 0.030763396993279457, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35930 + }, + { + "epoch": 54.045112781954884, + "grad_norm": 0.035094860941171646, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 35940 + }, + { + "epoch": 54.06015037593985, + "grad_norm": 0.017707301303744316, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35950 + }, + { + "epoch": 54.07518796992481, + "grad_norm": 0.03412231057882309, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 35960 + }, + { + "epoch": 54.090225563909776, + "grad_norm": 0.03027062863111496, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 35970 + }, + { + "epoch": 54.10526315789474, + "grad_norm": 0.014471354894340038, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 35980 + }, + { + "epoch": 54.1203007518797, + "grad_norm": 0.016598977148532867, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 35990 + }, + { + "epoch": 54.13533834586466, + "grad_norm": 0.03272229805588722, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36000 + }, + { + "epoch": 54.13533834586466, + "eval_cer": 0.0192366019756841, + "eval_loss": 0.1000109389424324, + "eval_runtime": 161.2707, + "eval_samples_per_second": 99.379, + "eval_steps_per_second": 0.781, + "eval_wer": 0.06912364449623185, + "step": 36000 + }, + { + "epoch": 54.150375939849624, + "grad_norm": 0.020665215328335762, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36010 + }, + { + "epoch": 54.16541353383459, + "grad_norm": 0.01658949814736843, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36020 + }, + { + "epoch": 54.18045112781955, + "grad_norm": 0.02667887695133686, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 36030 + }, + { + "epoch": 54.19548872180451, + "grad_norm": 0.03111160919070244, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36040 + }, + { + "epoch": 54.21052631578947, + "grad_norm": 0.023847274482250214, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 36050 + }, + { + "epoch": 54.225563909774436, + "grad_norm": 0.025221683084964752, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36060 + }, + { + "epoch": 54.2406015037594, + "grad_norm": 0.02251250669360161, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 36070 + }, + { + "epoch": 54.255639097744364, + "grad_norm": 0.024141931906342506, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36080 + }, + { + "epoch": 54.27067669172932, + "grad_norm": 0.02221955545246601, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36090 + }, + { + "epoch": 54.285714285714285, + "grad_norm": 0.023152465000748634, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36100 + }, + { + "epoch": 54.30075187969925, + "grad_norm": 0.01982812024652958, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36110 + }, + { + "epoch": 54.31578947368421, + "grad_norm": 0.01585562340915203, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36120 + }, + { + "epoch": 54.330827067669176, + "grad_norm": 0.017260851338505745, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36130 + }, + { + "epoch": 54.34586466165413, + "grad_norm": 0.04423464089632034, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36140 + }, + { + "epoch": 54.3609022556391, + "grad_norm": 0.021907851099967957, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36150 + }, + { + "epoch": 54.37593984962406, + "grad_norm": 0.019298365339636803, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 36160 + }, + { + "epoch": 54.390977443609025, + "grad_norm": 0.018214227631688118, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 36170 + }, + { + "epoch": 54.40601503759399, + "grad_norm": 0.022334251552820206, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36180 + }, + { + "epoch": 54.421052631578945, + "grad_norm": 0.023595578968524933, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36190 + }, + { + "epoch": 54.43609022556391, + "grad_norm": 0.022061672061681747, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 36200 + }, + { + "epoch": 54.45112781954887, + "grad_norm": 0.03340713679790497, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36210 + }, + { + "epoch": 54.46616541353384, + "grad_norm": 0.03255641832947731, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 36220 + }, + { + "epoch": 54.4812030075188, + "grad_norm": 0.018282905220985413, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36230 + }, + { + "epoch": 54.49624060150376, + "grad_norm": 0.032853610813617706, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36240 + }, + { + "epoch": 54.51127819548872, + "grad_norm": 0.021829942241311073, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36250 + }, + { + "epoch": 54.526315789473685, + "grad_norm": 0.014404632151126862, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36260 + }, + { + "epoch": 54.54135338345865, + "grad_norm": 0.022099580615758896, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36270 + }, + { + "epoch": 54.556390977443606, + "grad_norm": 0.016254520043730736, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 36280 + }, + { + "epoch": 54.57142857142857, + "grad_norm": 0.02206442505121231, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36290 + }, + { + "epoch": 54.58646616541353, + "grad_norm": 0.019399842247366905, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36300 + }, + { + "epoch": 54.6015037593985, + "grad_norm": 0.02703453041613102, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36310 + }, + { + "epoch": 54.61654135338346, + "grad_norm": 0.022649742662906647, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36320 + }, + { + "epoch": 54.63157894736842, + "grad_norm": 0.025307081639766693, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36330 + }, + { + "epoch": 54.64661654135338, + "grad_norm": 0.025691548362374306, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36340 + }, + { + "epoch": 54.661654135338345, + "grad_norm": 0.023014510050415993, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36350 + }, + { + "epoch": 54.67669172932331, + "grad_norm": 0.021506592631340027, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36360 + }, + { + "epoch": 54.69172932330827, + "grad_norm": 0.016283124685287476, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36370 + }, + { + "epoch": 54.70676691729323, + "grad_norm": 0.025756122544407845, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36380 + }, + { + "epoch": 54.721804511278194, + "grad_norm": 0.017057068645954132, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36390 + }, + { + "epoch": 54.73684210526316, + "grad_norm": 0.022225042805075645, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 36400 + }, + { + "epoch": 54.75187969924812, + "grad_norm": 0.05109328031539917, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 36410 + }, + { + "epoch": 54.766917293233085, + "grad_norm": 0.026894941926002502, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 36420 + }, + { + "epoch": 54.78195488721804, + "grad_norm": 0.02174895629286766, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36430 + }, + { + "epoch": 54.796992481203006, + "grad_norm": 0.022635988891124725, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36440 + }, + { + "epoch": 54.81203007518797, + "grad_norm": 0.024307209998369217, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36450 + }, + { + "epoch": 54.82706766917293, + "grad_norm": 0.021154968068003654, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36460 + }, + { + "epoch": 54.8421052631579, + "grad_norm": 0.021501153707504272, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36470 + }, + { + "epoch": 54.857142857142854, + "grad_norm": 0.018416155129671097, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36480 + }, + { + "epoch": 54.87218045112782, + "grad_norm": 0.02585945837199688, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36490 + }, + { + "epoch": 54.88721804511278, + "grad_norm": 0.01504368893802166, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36500 + }, + { + "epoch": 54.902255639097746, + "grad_norm": 0.017342112958431244, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36510 + }, + { + "epoch": 54.91729323308271, + "grad_norm": 0.01951522007584572, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36520 + }, + { + "epoch": 54.932330827067666, + "grad_norm": 0.02004878595471382, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36530 + }, + { + "epoch": 54.94736842105263, + "grad_norm": 0.043497566133737564, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36540 + }, + { + "epoch": 54.962406015037594, + "grad_norm": 0.028984608128666878, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36550 + }, + { + "epoch": 54.97744360902256, + "grad_norm": 0.017731167376041412, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 36560 + }, + { + "epoch": 54.99248120300752, + "grad_norm": 0.029077060520648956, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36570 + }, + { + "epoch": 55.00751879699248, + "grad_norm": 0.016465773805975914, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36580 + }, + { + "epoch": 55.02255639097744, + "grad_norm": 0.030533140525221825, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36590 + }, + { + "epoch": 55.037593984962406, + "grad_norm": 0.016080539673566818, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 36600 + }, + { + "epoch": 55.05263157894737, + "grad_norm": 0.02696414291858673, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36610 + }, + { + "epoch": 55.067669172932334, + "grad_norm": 0.02270100824534893, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 36620 + }, + { + "epoch": 55.08270676691729, + "grad_norm": 0.025448838248848915, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36630 + }, + { + "epoch": 55.097744360902254, + "grad_norm": 0.028867192566394806, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36640 + }, + { + "epoch": 55.11278195488722, + "grad_norm": 0.03476082906126976, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36650 + }, + { + "epoch": 55.12781954887218, + "grad_norm": 0.026140959933400154, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 36660 + }, + { + "epoch": 55.142857142857146, + "grad_norm": 0.016784565523266792, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36670 + }, + { + "epoch": 55.1578947368421, + "grad_norm": 0.0231111329048872, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 36680 + }, + { + "epoch": 55.17293233082707, + "grad_norm": 0.03465709090232849, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36690 + }, + { + "epoch": 55.18796992481203, + "grad_norm": 0.019499775022268295, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36700 + }, + { + "epoch": 55.203007518796994, + "grad_norm": 0.018681712448596954, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36710 + }, + { + "epoch": 55.21804511278196, + "grad_norm": 0.018785161897540092, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 36720 + }, + { + "epoch": 55.233082706766915, + "grad_norm": 0.043624408543109894, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36730 + }, + { + "epoch": 55.24812030075188, + "grad_norm": 0.02722357027232647, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36740 + }, + { + "epoch": 55.26315789473684, + "grad_norm": 0.0186101496219635, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36750 + }, + { + "epoch": 55.278195488721806, + "grad_norm": 0.014018344692885876, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36760 + }, + { + "epoch": 55.29323308270677, + "grad_norm": 0.021687373518943787, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36770 + }, + { + "epoch": 55.30827067669173, + "grad_norm": 0.01775556057691574, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 36780 + }, + { + "epoch": 55.32330827067669, + "grad_norm": 0.02582927793264389, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36790 + }, + { + "epoch": 55.338345864661655, + "grad_norm": 0.041416049003601074, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36800 + }, + { + "epoch": 55.35338345864662, + "grad_norm": 0.023935161530971527, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36810 + }, + { + "epoch": 55.36842105263158, + "grad_norm": 0.020043788477778435, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36820 + }, + { + "epoch": 55.38345864661654, + "grad_norm": 0.02871558628976345, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36830 + }, + { + "epoch": 55.3984962406015, + "grad_norm": 0.023433780297636986, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36840 + }, + { + "epoch": 55.41353383458647, + "grad_norm": 0.030400846153497696, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36850 + }, + { + "epoch": 55.42857142857143, + "grad_norm": 0.014550800435245037, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36860 + }, + { + "epoch": 55.443609022556394, + "grad_norm": 0.015704233199357986, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 36870 + }, + { + "epoch": 55.45864661654135, + "grad_norm": 0.02316436544060707, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 36880 + }, + { + "epoch": 55.473684210526315, + "grad_norm": 0.04076072946190834, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 36890 + }, + { + "epoch": 55.48872180451128, + "grad_norm": 0.03778422251343727, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36900 + }, + { + "epoch": 55.50375939849624, + "grad_norm": 0.01366632990539074, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 36910 + }, + { + "epoch": 55.5187969924812, + "grad_norm": 0.03566605597734451, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36920 + }, + { + "epoch": 55.53383458646616, + "grad_norm": 0.01601555570960045, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36930 + }, + { + "epoch": 55.54887218045113, + "grad_norm": 0.02527773194015026, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36940 + }, + { + "epoch": 55.56390977443609, + "grad_norm": 0.018258357420563698, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36950 + }, + { + "epoch": 55.578947368421055, + "grad_norm": 0.012750630266964436, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 36960 + }, + { + "epoch": 55.59398496240601, + "grad_norm": 0.03217349201440811, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36970 + }, + { + "epoch": 55.609022556390975, + "grad_norm": 0.027449045330286026, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 36980 + }, + { + "epoch": 55.62406015037594, + "grad_norm": 0.018417444080114365, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 36990 + }, + { + "epoch": 55.6390977443609, + "grad_norm": 0.025927415117621422, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37000 + }, + { + "epoch": 55.6390977443609, + "eval_cer": 0.01968339546732423, + "eval_loss": 0.1029498353600502, + "eval_runtime": 162.3792, + "eval_samples_per_second": 98.701, + "eval_steps_per_second": 0.776, + "eval_wer": 0.06809938168354977, + "step": 37000 + }, + { + "epoch": 55.65413533834587, + "grad_norm": 0.030380457639694214, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37010 + }, + { + "epoch": 55.669172932330824, + "grad_norm": 0.02089093066751957, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 37020 + }, + { + "epoch": 55.68421052631579, + "grad_norm": 0.019579371437430382, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37030 + }, + { + "epoch": 55.69924812030075, + "grad_norm": 0.025974644348025322, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37040 + }, + { + "epoch": 55.714285714285715, + "grad_norm": 0.02742829918861389, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37050 + }, + { + "epoch": 55.72932330827068, + "grad_norm": 0.025304416194558144, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37060 + }, + { + "epoch": 55.744360902255636, + "grad_norm": 0.02115757390856743, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37070 + }, + { + "epoch": 55.7593984962406, + "grad_norm": 0.026871832087635994, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37080 + }, + { + "epoch": 55.774436090225564, + "grad_norm": 0.03919268399477005, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 37090 + }, + { + "epoch": 55.78947368421053, + "grad_norm": 0.020050985738635063, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37100 + }, + { + "epoch": 55.80451127819549, + "grad_norm": 0.034043584018945694, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37110 + }, + { + "epoch": 55.81954887218045, + "grad_norm": 0.013206139206886292, + "learning_rate": 0.0005, + "loss": 0.0018, + "step": 37120 + }, + { + "epoch": 55.83458646616541, + "grad_norm": 0.022174932062625885, + "learning_rate": 0.0005, + "loss": 0.0022, + "step": 37130 + }, + { + "epoch": 55.849624060150376, + "grad_norm": 0.017640048637986183, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37140 + }, + { + "epoch": 55.86466165413534, + "grad_norm": 0.02204502560198307, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37150 + }, + { + "epoch": 55.8796992481203, + "grad_norm": 0.019784940406680107, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37160 + }, + { + "epoch": 55.89473684210526, + "grad_norm": 0.02361452765762806, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37170 + }, + { + "epoch": 55.909774436090224, + "grad_norm": 0.03177280351519585, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37180 + }, + { + "epoch": 55.92481203007519, + "grad_norm": 0.032261431217193604, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37190 + }, + { + "epoch": 55.93984962406015, + "grad_norm": 0.018549378961324692, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37200 + }, + { + "epoch": 55.954887218045116, + "grad_norm": 0.02213878370821476, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37210 + }, + { + "epoch": 55.96992481203007, + "grad_norm": 0.02742697484791279, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37220 + }, + { + "epoch": 55.984962406015036, + "grad_norm": 0.02579023689031601, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37230 + }, + { + "epoch": 56.0, + "grad_norm": 0.024157574400305748, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37240 + }, + { + "epoch": 56.015037593984964, + "grad_norm": 0.01995805650949478, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37250 + }, + { + "epoch": 56.03007518796993, + "grad_norm": 0.019094256684184074, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37260 + }, + { + "epoch": 56.045112781954884, + "grad_norm": 0.01880325749516487, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37270 + }, + { + "epoch": 56.06015037593985, + "grad_norm": 0.018590591847896576, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37280 + }, + { + "epoch": 56.07518796992481, + "grad_norm": 0.01915215328335762, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37290 + }, + { + "epoch": 56.090225563909776, + "grad_norm": 0.02245292067527771, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37300 + }, + { + "epoch": 56.10526315789474, + "grad_norm": 0.042971763759851456, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37310 + }, + { + "epoch": 56.1203007518797, + "grad_norm": 0.029251370579004288, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 37320 + }, + { + "epoch": 56.13533834586466, + "grad_norm": 0.019463282078504562, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37330 + }, + { + "epoch": 56.150375939849624, + "grad_norm": 0.018303487449884415, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37340 + }, + { + "epoch": 56.16541353383459, + "grad_norm": 0.016320591792464256, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37350 + }, + { + "epoch": 56.18045112781955, + "grad_norm": 0.02417987398803234, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 37360 + }, + { + "epoch": 56.19548872180451, + "grad_norm": 0.030176015570759773, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37370 + }, + { + "epoch": 56.21052631578947, + "grad_norm": 0.025113999843597412, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37380 + }, + { + "epoch": 56.225563909774436, + "grad_norm": 0.02251431904733181, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37390 + }, + { + "epoch": 56.2406015037594, + "grad_norm": 0.02240750752389431, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37400 + }, + { + "epoch": 56.255639097744364, + "grad_norm": 0.027555342763662338, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37410 + }, + { + "epoch": 56.27067669172932, + "grad_norm": 0.022266382351517677, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37420 + }, + { + "epoch": 56.285714285714285, + "grad_norm": 0.022692464292049408, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37430 + }, + { + "epoch": 56.30075187969925, + "grad_norm": 0.021963071078062057, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37440 + }, + { + "epoch": 56.31578947368421, + "grad_norm": 0.008833964355289936, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 37450 + }, + { + "epoch": 56.330827067669176, + "grad_norm": 0.015709925442934036, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37460 + }, + { + "epoch": 56.34586466165413, + "grad_norm": 0.015945546329021454, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37470 + }, + { + "epoch": 56.3609022556391, + "grad_norm": 0.021570047363638878, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37480 + }, + { + "epoch": 56.37593984962406, + "grad_norm": 0.02429301105439663, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37490 + }, + { + "epoch": 56.390977443609025, + "grad_norm": 0.025561930611729622, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37500 + }, + { + "epoch": 56.40601503759399, + "grad_norm": 0.02860790118575096, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37510 + }, + { + "epoch": 56.421052631578945, + "grad_norm": 0.031526707112789154, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37520 + }, + { + "epoch": 56.43609022556391, + "grad_norm": 0.01981993578374386, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37530 + }, + { + "epoch": 56.45112781954887, + "grad_norm": 0.041662294417619705, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37540 + }, + { + "epoch": 56.46616541353384, + "grad_norm": 0.018954534083604813, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37550 + }, + { + "epoch": 56.4812030075188, + "grad_norm": 0.019187292084097862, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37560 + }, + { + "epoch": 56.49624060150376, + "grad_norm": 0.0174203272908926, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37570 + }, + { + "epoch": 56.51127819548872, + "grad_norm": 0.015347965992987156, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37580 + }, + { + "epoch": 56.526315789473685, + "grad_norm": 0.022027455270290375, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37590 + }, + { + "epoch": 56.54135338345865, + "grad_norm": 0.018741143867373466, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 37600 + }, + { + "epoch": 56.556390977443606, + "grad_norm": 0.03647798299789429, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37610 + }, + { + "epoch": 56.57142857142857, + "grad_norm": 0.017190013080835342, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37620 + }, + { + "epoch": 56.58646616541353, + "grad_norm": 0.04564463719725609, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37630 + }, + { + "epoch": 56.6015037593985, + "grad_norm": 0.02394207939505577, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37640 + }, + { + "epoch": 56.61654135338346, + "grad_norm": 0.021602340042591095, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37650 + }, + { + "epoch": 56.63157894736842, + "grad_norm": 0.013426977209746838, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37660 + }, + { + "epoch": 56.64661654135338, + "grad_norm": 0.013105706311762333, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37670 + }, + { + "epoch": 56.661654135338345, + "grad_norm": 0.0182881448417902, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37680 + }, + { + "epoch": 56.67669172932331, + "grad_norm": 0.019317694008350372, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37690 + }, + { + "epoch": 56.69172932330827, + "grad_norm": 0.011145003139972687, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 37700 + }, + { + "epoch": 56.70676691729323, + "grad_norm": 0.017270414158701897, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37710 + }, + { + "epoch": 56.721804511278194, + "grad_norm": 0.02993171103298664, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37720 + }, + { + "epoch": 56.73684210526316, + "grad_norm": 0.021182319149374962, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37730 + }, + { + "epoch": 56.75187969924812, + "grad_norm": 0.020071232691407204, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37740 + }, + { + "epoch": 56.766917293233085, + "grad_norm": 0.021860240027308464, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37750 + }, + { + "epoch": 56.78195488721804, + "grad_norm": 0.07659654319286346, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 37760 + }, + { + "epoch": 56.796992481203006, + "grad_norm": 0.0209948867559433, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37770 + }, + { + "epoch": 56.81203007518797, + "grad_norm": 0.029444528743624687, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37780 + }, + { + "epoch": 56.82706766917293, + "grad_norm": 0.026411548256874084, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37790 + }, + { + "epoch": 56.8421052631579, + "grad_norm": 0.023158356547355652, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37800 + }, + { + "epoch": 56.857142857142854, + "grad_norm": 0.023941583931446075, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 37810 + }, + { + "epoch": 56.87218045112782, + "grad_norm": 0.04632248729467392, + "learning_rate": 0.0005, + "loss": 0.0019, + "step": 37820 + }, + { + "epoch": 56.88721804511278, + "grad_norm": 0.021189412102103233, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37830 + }, + { + "epoch": 56.902255639097746, + "grad_norm": 0.02353930100798607, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37840 + }, + { + "epoch": 56.91729323308271, + "grad_norm": 0.019352585077285767, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 37850 + }, + { + "epoch": 56.932330827067666, + "grad_norm": 0.017305372282862663, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37860 + }, + { + "epoch": 56.94736842105263, + "grad_norm": 0.023371415212750435, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37870 + }, + { + "epoch": 56.962406015037594, + "grad_norm": 0.026973599568009377, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37880 + }, + { + "epoch": 56.97744360902256, + "grad_norm": 0.032823625952005386, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37890 + }, + { + "epoch": 56.99248120300752, + "grad_norm": 0.038597408682107925, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37900 + }, + { + "epoch": 57.00751879699248, + "grad_norm": 0.02546747401356697, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 37910 + }, + { + "epoch": 57.02255639097744, + "grad_norm": 0.01972416415810585, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37920 + }, + { + "epoch": 57.037593984962406, + "grad_norm": 0.02202797681093216, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37930 + }, + { + "epoch": 57.05263157894737, + "grad_norm": 0.024695521220564842, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37940 + }, + { + "epoch": 57.067669172932334, + "grad_norm": 0.020123912021517754, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37950 + }, + { + "epoch": 57.08270676691729, + "grad_norm": 0.022145427763462067, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 37960 + }, + { + "epoch": 57.097744360902254, + "grad_norm": 0.018137359991669655, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37970 + }, + { + "epoch": 57.11278195488722, + "grad_norm": 0.01721174828708172, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 37980 + }, + { + "epoch": 57.12781954887218, + "grad_norm": 0.020454922690987587, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 37990 + }, + { + "epoch": 57.142857142857146, + "grad_norm": 0.02048588916659355, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38000 + }, + { + "epoch": 57.142857142857146, + "eval_cer": 0.019240002230567203, + "eval_loss": 0.10025910288095474, + "eval_runtime": 161.6297, + "eval_samples_per_second": 99.159, + "eval_steps_per_second": 0.78, + "eval_wer": 0.0684423687722002, + "step": 38000 + }, + { + "epoch": 57.1578947368421, + "grad_norm": 0.014652017503976822, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38010 + }, + { + "epoch": 57.17293233082707, + "grad_norm": 0.014320539310574532, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38020 + }, + { + "epoch": 57.18796992481203, + "grad_norm": 0.01915593259036541, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38030 + }, + { + "epoch": 57.203007518796994, + "grad_norm": 0.022277556359767914, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38040 + }, + { + "epoch": 57.21804511278196, + "grad_norm": 0.01937675103545189, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 38050 + }, + { + "epoch": 57.233082706766915, + "grad_norm": 0.015909263864159584, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38060 + }, + { + "epoch": 57.24812030075188, + "grad_norm": 0.021855473518371582, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38070 + }, + { + "epoch": 57.26315789473684, + "grad_norm": 0.02376752533018589, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38080 + }, + { + "epoch": 57.278195488721806, + "grad_norm": 0.02349868044257164, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38090 + }, + { + "epoch": 57.29323308270677, + "grad_norm": 0.025749806314706802, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38100 + }, + { + "epoch": 57.30827067669173, + "grad_norm": 0.03801216185092926, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38110 + }, + { + "epoch": 57.32330827067669, + "grad_norm": 0.034677986055612564, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38120 + }, + { + "epoch": 57.338345864661655, + "grad_norm": 0.01573493331670761, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38130 + }, + { + "epoch": 57.35338345864662, + "grad_norm": 0.016923772171139717, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38140 + }, + { + "epoch": 57.36842105263158, + "grad_norm": 0.024984369054436684, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38150 + }, + { + "epoch": 57.38345864661654, + "grad_norm": 0.018282465636730194, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38160 + }, + { + "epoch": 57.3984962406015, + "grad_norm": 0.028648560866713524, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38170 + }, + { + "epoch": 57.41353383458647, + "grad_norm": 0.018610531464219093, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38180 + }, + { + "epoch": 57.42857142857143, + "grad_norm": 0.023726286366581917, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38190 + }, + { + "epoch": 57.443609022556394, + "grad_norm": 0.022759221494197845, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38200 + }, + { + "epoch": 57.45864661654135, + "grad_norm": 0.021196119487285614, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38210 + }, + { + "epoch": 57.473684210526315, + "grad_norm": 0.015172051265835762, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38220 + }, + { + "epoch": 57.48872180451128, + "grad_norm": 0.024735068902373314, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38230 + }, + { + "epoch": 57.50375939849624, + "grad_norm": 0.019010029733181, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38240 + }, + { + "epoch": 57.5187969924812, + "grad_norm": 0.02551075629889965, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38250 + }, + { + "epoch": 57.53383458646616, + "grad_norm": 0.02254568599164486, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38260 + }, + { + "epoch": 57.54887218045113, + "grad_norm": 0.02857019007205963, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38270 + }, + { + "epoch": 57.56390977443609, + "grad_norm": 0.017542444169521332, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38280 + }, + { + "epoch": 57.578947368421055, + "grad_norm": 0.026463152840733528, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38290 + }, + { + "epoch": 57.59398496240601, + "grad_norm": 0.02033921517431736, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38300 + }, + { + "epoch": 57.609022556390975, + "grad_norm": 0.026763344183564186, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38310 + }, + { + "epoch": 57.62406015037594, + "grad_norm": 0.033811140805482864, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38320 + }, + { + "epoch": 57.6390977443609, + "grad_norm": 0.02197147160768509, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38330 + }, + { + "epoch": 57.65413533834587, + "grad_norm": 0.02602764591574669, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38340 + }, + { + "epoch": 57.669172932330824, + "grad_norm": 0.022421633824706078, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38350 + }, + { + "epoch": 57.68421052631579, + "grad_norm": 0.010614289902150631, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38360 + }, + { + "epoch": 57.69924812030075, + "grad_norm": 0.018742339685559273, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38370 + }, + { + "epoch": 57.714285714285715, + "grad_norm": 0.03064779005944729, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38380 + }, + { + "epoch": 57.72932330827068, + "grad_norm": 0.018140144646167755, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38390 + }, + { + "epoch": 57.744360902255636, + "grad_norm": 0.025918951258063316, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 38400 + }, + { + "epoch": 57.7593984962406, + "grad_norm": 0.021767975762486458, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38410 + }, + { + "epoch": 57.774436090225564, + "grad_norm": 0.023072658106684685, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38420 + }, + { + "epoch": 57.78947368421053, + "grad_norm": 0.025459900498390198, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38430 + }, + { + "epoch": 57.80451127819549, + "grad_norm": 0.02249845117330551, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38440 + }, + { + "epoch": 57.81954887218045, + "grad_norm": 0.03887522593140602, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38450 + }, + { + "epoch": 57.83458646616541, + "grad_norm": 0.02365991845726967, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38460 + }, + { + "epoch": 57.849624060150376, + "grad_norm": 0.03075840324163437, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38470 + }, + { + "epoch": 57.86466165413534, + "grad_norm": 0.02918490208685398, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38480 + }, + { + "epoch": 57.8796992481203, + "grad_norm": 0.020126987248659134, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38490 + }, + { + "epoch": 57.89473684210526, + "grad_norm": 0.018086997792124748, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38500 + }, + { + "epoch": 57.909774436090224, + "grad_norm": 0.02490510232746601, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38510 + }, + { + "epoch": 57.92481203007519, + "grad_norm": 0.028850378468632698, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38520 + }, + { + "epoch": 57.93984962406015, + "grad_norm": 0.02102653682231903, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38530 + }, + { + "epoch": 57.954887218045116, + "grad_norm": 0.008839157409965992, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38540 + }, + { + "epoch": 57.96992481203007, + "grad_norm": 0.01821914315223694, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38550 + }, + { + "epoch": 57.984962406015036, + "grad_norm": 0.02789415791630745, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38560 + }, + { + "epoch": 58.0, + "grad_norm": 0.010872890241444111, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38570 + }, + { + "epoch": 58.015037593984964, + "grad_norm": 0.01939144916832447, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38580 + }, + { + "epoch": 58.03007518796993, + "grad_norm": 0.013604406267404556, + "learning_rate": 0.0005, + "loss": 0.0008, + "step": 38590 + }, + { + "epoch": 58.045112781954884, + "grad_norm": 0.022815557196736336, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38600 + }, + { + "epoch": 58.06015037593985, + "grad_norm": 0.017643490806221962, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38610 + }, + { + "epoch": 58.07518796992481, + "grad_norm": 0.02417610026896, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 38620 + }, + { + "epoch": 58.090225563909776, + "grad_norm": 0.024080639705061913, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 38630 + }, + { + "epoch": 58.10526315789474, + "grad_norm": 0.03251300007104874, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38640 + }, + { + "epoch": 58.1203007518797, + "grad_norm": 0.024090873077511787, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38650 + }, + { + "epoch": 58.13533834586466, + "grad_norm": 0.0182518120855093, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38660 + }, + { + "epoch": 58.150375939849624, + "grad_norm": 0.019323313608765602, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38670 + }, + { + "epoch": 58.16541353383459, + "grad_norm": 0.015427827835083008, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38680 + }, + { + "epoch": 58.18045112781955, + "grad_norm": 0.016766540706157684, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38690 + }, + { + "epoch": 58.19548872180451, + "grad_norm": 0.03185683488845825, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38700 + }, + { + "epoch": 58.21052631578947, + "grad_norm": 0.020946508273482323, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38710 + }, + { + "epoch": 58.225563909774436, + "grad_norm": 0.018846899271011353, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38720 + }, + { + "epoch": 58.2406015037594, + "grad_norm": 0.029208123683929443, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38730 + }, + { + "epoch": 58.255639097744364, + "grad_norm": 0.015121311880648136, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38740 + }, + { + "epoch": 58.27067669172932, + "grad_norm": 0.014629244804382324, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38750 + }, + { + "epoch": 58.285714285714285, + "grad_norm": 0.033221881836652756, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38760 + }, + { + "epoch": 58.30075187969925, + "grad_norm": 0.014489257708191872, + "learning_rate": 0.0005, + "loss": 0.0016, + "step": 38770 + }, + { + "epoch": 58.31578947368421, + "grad_norm": 0.02097286470234394, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38780 + }, + { + "epoch": 58.330827067669176, + "grad_norm": 0.029289226979017258, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38790 + }, + { + "epoch": 58.34586466165413, + "grad_norm": 0.024109337478876114, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38800 + }, + { + "epoch": 58.3609022556391, + "grad_norm": 0.020232167094945908, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38810 + }, + { + "epoch": 58.37593984962406, + "grad_norm": 0.018179412931203842, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 38820 + }, + { + "epoch": 58.390977443609025, + "grad_norm": 0.015574472025036812, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38830 + }, + { + "epoch": 58.40601503759399, + "grad_norm": 0.03568084165453911, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38840 + }, + { + "epoch": 58.421052631578945, + "grad_norm": 0.018514027819037437, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38850 + }, + { + "epoch": 58.43609022556391, + "grad_norm": 0.04555121809244156, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38860 + }, + { + "epoch": 58.45112781954887, + "grad_norm": 0.019877567887306213, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38870 + }, + { + "epoch": 58.46616541353384, + "grad_norm": 0.016229791566729546, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38880 + }, + { + "epoch": 58.4812030075188, + "grad_norm": 0.04647158458828926, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38890 + }, + { + "epoch": 58.49624060150376, + "grad_norm": 0.02335761860013008, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38900 + }, + { + "epoch": 58.51127819548872, + "grad_norm": 0.01706009916961193, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38910 + }, + { + "epoch": 58.526315789473685, + "grad_norm": 0.02373575046658516, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38920 + }, + { + "epoch": 58.54135338345865, + "grad_norm": 0.01922602206468582, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 38930 + }, + { + "epoch": 58.556390977443606, + "grad_norm": 0.026201875880360603, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 38940 + }, + { + "epoch": 58.57142857142857, + "grad_norm": 0.025738239288330078, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38950 + }, + { + "epoch": 58.58646616541353, + "grad_norm": 0.02361173927783966, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 38960 + }, + { + "epoch": 58.6015037593985, + "grad_norm": 0.01953188329935074, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38970 + }, + { + "epoch": 58.61654135338346, + "grad_norm": 0.016951201483607292, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 38980 + }, + { + "epoch": 58.63157894736842, + "grad_norm": 0.030946874991059303, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 38990 + }, + { + "epoch": 58.64661654135338, + "grad_norm": 0.018709134310483932, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39000 + }, + { + "epoch": 58.64661654135338, + "eval_cer": 0.019316167939948777, + "eval_loss": 0.0985691249370575, + "eval_runtime": 161.1351, + "eval_samples_per_second": 99.463, + "eval_steps_per_second": 0.782, + "eval_wer": 0.06763893326316976, + "step": 39000 + }, + { + "epoch": 58.661654135338345, + "grad_norm": 0.02820403501391411, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39010 + }, + { + "epoch": 58.67669172932331, + "grad_norm": 0.021069442853331566, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39020 + }, + { + "epoch": 58.69172932330827, + "grad_norm": 0.019026735797524452, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39030 + }, + { + "epoch": 58.70676691729323, + "grad_norm": 0.01603548787534237, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39040 + }, + { + "epoch": 58.721804511278194, + "grad_norm": 0.022312704473733902, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39050 + }, + { + "epoch": 58.73684210526316, + "grad_norm": 0.02293722704052925, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39060 + }, + { + "epoch": 58.75187969924812, + "grad_norm": 0.015065363608300686, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39070 + }, + { + "epoch": 58.766917293233085, + "grad_norm": 0.022291777655482292, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39080 + }, + { + "epoch": 58.78195488721804, + "grad_norm": 0.02543407678604126, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39090 + }, + { + "epoch": 58.796992481203006, + "grad_norm": 0.021084189414978027, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39100 + }, + { + "epoch": 58.81203007518797, + "grad_norm": 0.037318117916584015, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39110 + }, + { + "epoch": 58.82706766917293, + "grad_norm": 0.0349205806851387, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39120 + }, + { + "epoch": 58.8421052631579, + "grad_norm": 0.022080961614847183, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39130 + }, + { + "epoch": 58.857142857142854, + "grad_norm": 0.025429993867874146, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39140 + }, + { + "epoch": 58.87218045112782, + "grad_norm": 0.026794854551553726, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39150 + }, + { + "epoch": 58.88721804511278, + "grad_norm": 0.014020625501871109, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39160 + }, + { + "epoch": 58.902255639097746, + "grad_norm": 0.021101457998156548, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39170 + }, + { + "epoch": 58.91729323308271, + "grad_norm": 0.03400740027427673, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39180 + }, + { + "epoch": 58.932330827067666, + "grad_norm": 0.03134303539991379, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39190 + }, + { + "epoch": 58.94736842105263, + "grad_norm": 0.03083951584994793, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39200 + }, + { + "epoch": 58.962406015037594, + "grad_norm": 0.017887147143483162, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39210 + }, + { + "epoch": 58.97744360902256, + "grad_norm": 0.020240947604179382, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39220 + }, + { + "epoch": 58.99248120300752, + "grad_norm": 0.02022034116089344, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39230 + }, + { + "epoch": 59.00751879699248, + "grad_norm": 0.031057924032211304, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39240 + }, + { + "epoch": 59.02255639097744, + "grad_norm": 0.030442899093031883, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39250 + }, + { + "epoch": 59.037593984962406, + "grad_norm": 0.018583735451102257, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39260 + }, + { + "epoch": 59.05263157894737, + "grad_norm": 0.019672812893986702, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39270 + }, + { + "epoch": 59.067669172932334, + "grad_norm": 0.016705363988876343, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39280 + }, + { + "epoch": 59.08270676691729, + "grad_norm": 0.04884115606546402, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 39290 + }, + { + "epoch": 59.097744360902254, + "grad_norm": 0.01807328127324581, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39300 + }, + { + "epoch": 59.11278195488722, + "grad_norm": 0.020648999139666557, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39310 + }, + { + "epoch": 59.12781954887218, + "grad_norm": 0.02000991813838482, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39320 + }, + { + "epoch": 59.142857142857146, + "grad_norm": 0.02096889168024063, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39330 + }, + { + "epoch": 59.1578947368421, + "grad_norm": 0.022525062784552574, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39340 + }, + { + "epoch": 59.17293233082707, + "grad_norm": 0.03318751975893974, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39350 + }, + { + "epoch": 59.18796992481203, + "grad_norm": 0.0245821513235569, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39360 + }, + { + "epoch": 59.203007518796994, + "grad_norm": 0.02982010692358017, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39370 + }, + { + "epoch": 59.21804511278196, + "grad_norm": 0.029251564294099808, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39380 + }, + { + "epoch": 59.233082706766915, + "grad_norm": 0.02026515081524849, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39390 + }, + { + "epoch": 59.24812030075188, + "grad_norm": 0.027000311762094498, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39400 + }, + { + "epoch": 59.26315789473684, + "grad_norm": 0.01281500793993473, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39410 + }, + { + "epoch": 59.278195488721806, + "grad_norm": 0.03465868905186653, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39420 + }, + { + "epoch": 59.29323308270677, + "grad_norm": 0.02127237804234028, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39430 + }, + { + "epoch": 59.30827067669173, + "grad_norm": 0.021957850083708763, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39440 + }, + { + "epoch": 59.32330827067669, + "grad_norm": 0.019679982215166092, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39450 + }, + { + "epoch": 59.338345864661655, + "grad_norm": 0.021086590364575386, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39460 + }, + { + "epoch": 59.35338345864662, + "grad_norm": 0.016761232167482376, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39470 + }, + { + "epoch": 59.36842105263158, + "grad_norm": 0.021811911836266518, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39480 + }, + { + "epoch": 59.38345864661654, + "grad_norm": 0.022179067134857178, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39490 + }, + { + "epoch": 59.3984962406015, + "grad_norm": 0.02547740936279297, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39500 + }, + { + "epoch": 59.41353383458647, + "grad_norm": 0.034297145903110504, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39510 + }, + { + "epoch": 59.42857142857143, + "grad_norm": 0.03346757963299751, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39520 + }, + { + "epoch": 59.443609022556394, + "grad_norm": 0.021065112203359604, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39530 + }, + { + "epoch": 59.45864661654135, + "grad_norm": 0.018498249351978302, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39540 + }, + { + "epoch": 59.473684210526315, + "grad_norm": 0.03482156619429588, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39550 + }, + { + "epoch": 59.48872180451128, + "grad_norm": 0.019200988113880157, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39560 + }, + { + "epoch": 59.50375939849624, + "grad_norm": 0.02717105858027935, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39570 + }, + { + "epoch": 59.5187969924812, + "grad_norm": 0.022116724401712418, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39580 + }, + { + "epoch": 59.53383458646616, + "grad_norm": 0.023991255089640617, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39590 + }, + { + "epoch": 59.54887218045113, + "grad_norm": 0.011700261384248734, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39600 + }, + { + "epoch": 59.56390977443609, + "grad_norm": 0.02005104348063469, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39610 + }, + { + "epoch": 59.578947368421055, + "grad_norm": 0.02267971634864807, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39620 + }, + { + "epoch": 59.59398496240601, + "grad_norm": 0.020618146285414696, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39630 + }, + { + "epoch": 59.609022556390975, + "grad_norm": 0.016169648617506027, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39640 + }, + { + "epoch": 59.62406015037594, + "grad_norm": 0.0288933627307415, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39650 + }, + { + "epoch": 59.6390977443609, + "grad_norm": 0.02188965678215027, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39660 + }, + { + "epoch": 59.65413533834587, + "grad_norm": 0.025946790352463722, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39670 + }, + { + "epoch": 59.669172932330824, + "grad_norm": 0.02399466186761856, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39680 + }, + { + "epoch": 59.68421052631579, + "grad_norm": 0.014388099312782288, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39690 + }, + { + "epoch": 59.69924812030075, + "grad_norm": 0.014940574765205383, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39700 + }, + { + "epoch": 59.714285714285715, + "grad_norm": 0.016061697155237198, + "learning_rate": 0.0005, + "loss": 0.0008, + "step": 39710 + }, + { + "epoch": 59.72932330827068, + "grad_norm": 0.02132270857691765, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39720 + }, + { + "epoch": 59.744360902255636, + "grad_norm": 0.023961182683706284, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39730 + }, + { + "epoch": 59.7593984962406, + "grad_norm": 0.02142259292304516, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39740 + }, + { + "epoch": 59.774436090225564, + "grad_norm": 0.022281289100646973, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39750 + }, + { + "epoch": 59.78947368421053, + "grad_norm": 0.025485748425126076, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39760 + }, + { + "epoch": 59.80451127819549, + "grad_norm": 0.012654361315071583, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39770 + }, + { + "epoch": 59.81954887218045, + "grad_norm": 0.01137780025601387, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39780 + }, + { + "epoch": 59.83458646616541, + "grad_norm": 0.028939247131347656, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39790 + }, + { + "epoch": 59.849624060150376, + "grad_norm": 0.03210778906941414, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39800 + }, + { + "epoch": 59.86466165413534, + "grad_norm": 0.029495107010006905, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 39810 + }, + { + "epoch": 59.8796992481203, + "grad_norm": 0.03220450505614281, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39820 + }, + { + "epoch": 59.89473684210526, + "grad_norm": 0.025627823546528816, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39830 + }, + { + "epoch": 59.909774436090224, + "grad_norm": 0.027897870168089867, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39840 + }, + { + "epoch": 59.92481203007519, + "grad_norm": 0.02983030490577221, + "learning_rate": 0.0005, + "loss": 0.0017, + "step": 39850 + }, + { + "epoch": 59.93984962406015, + "grad_norm": 0.01306893303990364, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39860 + }, + { + "epoch": 59.954887218045116, + "grad_norm": 0.03126494958996773, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39870 + }, + { + "epoch": 59.96992481203007, + "grad_norm": 0.02328731305897236, + "learning_rate": 0.0005, + "loss": 0.0013, + "step": 39880 + }, + { + "epoch": 59.984962406015036, + "grad_norm": 0.03705844655632973, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39890 + }, + { + "epoch": 60.0, + "grad_norm": 0.024091210216283798, + "learning_rate": 0.0005, + "loss": 0.0015, + "step": 39900 + }, + { + "epoch": 60.015037593984964, + "grad_norm": 0.0222784336656332, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39910 + }, + { + "epoch": 60.03007518796993, + "grad_norm": 0.025883842259645462, + "learning_rate": 0.0005, + "loss": 0.0014, + "step": 39920 + }, + { + "epoch": 60.045112781954884, + "grad_norm": 0.020213134586811066, + "learning_rate": 0.0005, + "loss": 0.0009, + "step": 39930 + }, + { + "epoch": 60.06015037593985, + "grad_norm": 0.01888001337647438, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39940 + }, + { + "epoch": 60.07518796992481, + "grad_norm": 0.02916104905307293, + "learning_rate": 0.0005, + "loss": 0.0011, + "step": 39950 + }, + { + "epoch": 60.090225563909776, + "grad_norm": 0.030693048611283302, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39960 + }, + { + "epoch": 60.10526315789474, + "grad_norm": 0.019310826435685158, + "learning_rate": 0.0005, + "loss": 0.0012, + "step": 39970 + }, + { + "epoch": 60.1203007518797, + "grad_norm": 0.012031104415655136, + "learning_rate": 0.0005, + "loss": 0.0008, + "step": 39980 + }, + { + "epoch": 60.13533834586466, + "grad_norm": 0.030276013538241386, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 39990 + }, + { + "epoch": 60.150375939849624, + "grad_norm": 0.022855514660477638, + "learning_rate": 0.0005, + "loss": 0.001, + "step": 40000 + }, + { + "epoch": 60.150375939849624, + "eval_cer": 0.019533784252467566, + "eval_loss": 0.10252687335014343, + "eval_runtime": 162.5503, + "eval_samples_per_second": 98.597, + "eval_steps_per_second": 0.775, + "eval_wer": 0.07112518558890413, + "step": 40000 + }, + { + "epoch": 60.16541353383459, + "grad_norm": 0.014998081140220165, + "learning_rate": 0.0004999987663004646, + "loss": 0.001, + "step": 40010 + }, + { + "epoch": 60.18045112781955, + "grad_norm": 0.02979385107755661, + "learning_rate": 0.0004999950652140343, + "loss": 0.0013, + "step": 40020 + }, + { + "epoch": 60.19548872180451, + "grad_norm": 0.02879681997001171, + "learning_rate": 0.0004999888967772375, + "loss": 0.0013, + "step": 40030 + }, + { + "epoch": 60.21052631578947, + "grad_norm": 0.01753871701657772, + "learning_rate": 0.0004999802610509541, + "loss": 0.0012, + "step": 40040 + }, + { + "epoch": 60.225563909774436, + "grad_norm": 0.012824436649680138, + "learning_rate": 0.0004999691581204152, + "loss": 0.0014, + "step": 40050 + }, + { + "epoch": 60.2406015037594, + "grad_norm": 0.01807374507188797, + "learning_rate": 0.0004999555880952023, + "loss": 0.0013, + "step": 40060 + }, + { + "epoch": 60.255639097744364, + "grad_norm": 0.016114136204123497, + "learning_rate": 0.000499939551109246, + "loss": 0.0009, + "step": 40070 + }, + { + "epoch": 60.27067669172932, + "grad_norm": 0.01620345190167427, + "learning_rate": 0.000499921047320825, + "loss": 0.0013, + "step": 40080 + }, + { + "epoch": 60.285714285714285, + "grad_norm": 0.0180693157017231, + "learning_rate": 0.0004999000769125642, + "loss": 0.0013, + "step": 40090 + }, + { + "epoch": 60.30075187969925, + "grad_norm": 0.01564876176416874, + "learning_rate": 0.0004998766400914329, + "loss": 0.0013, + "step": 40100 + }, + { + "epoch": 60.31578947368421, + "grad_norm": 0.015593121759593487, + "learning_rate": 0.0004998507370887433, + "loss": 0.0012, + "step": 40110 + }, + { + "epoch": 60.330827067669176, + "grad_norm": 0.013081851415336132, + "learning_rate": 0.0004998223681601474, + "loss": 0.0009, + "step": 40120 + }, + { + "epoch": 60.34586466165413, + "grad_norm": 0.03683178126811981, + "learning_rate": 0.000499791533585635, + "loss": 0.0012, + "step": 40130 + }, + { + "epoch": 60.3609022556391, + "grad_norm": 0.023180559277534485, + "learning_rate": 0.0004997582336695312, + "loss": 0.001, + "step": 40140 + }, + { + "epoch": 60.37593984962406, + "grad_norm": 0.022562697529792786, + "learning_rate": 0.0004997224687404926, + "loss": 0.0014, + "step": 40150 + }, + { + "epoch": 60.390977443609025, + "grad_norm": 0.02034986950457096, + "learning_rate": 0.0004996842391515044, + "loss": 0.0012, + "step": 40160 + }, + { + "epoch": 60.40601503759399, + "grad_norm": 0.026195017620921135, + "learning_rate": 0.0004996435452798775, + "loss": 0.0015, + "step": 40170 + }, + { + "epoch": 60.421052631578945, + "grad_norm": 0.019972871989011765, + "learning_rate": 0.0004996003875272438, + "loss": 0.0009, + "step": 40180 + }, + { + "epoch": 60.43609022556391, + "grad_norm": 0.031110072508454323, + "learning_rate": 0.000499554766319553, + "loss": 0.0015, + "step": 40190 + }, + { + "epoch": 60.45112781954887, + "grad_norm": 0.01594257913529873, + "learning_rate": 0.0004995066821070679, + "loss": 0.0009, + "step": 40200 + }, + { + "epoch": 60.46616541353384, + "grad_norm": 0.02081778459250927, + "learning_rate": 0.0004994561353643604, + "loss": 0.0012, + "step": 40210 + }, + { + "epoch": 60.4812030075188, + "grad_norm": 0.020923923701047897, + "learning_rate": 0.0004994031265903063, + "loss": 0.0011, + "step": 40220 + }, + { + "epoch": 60.49624060150376, + "grad_norm": 0.017224030569195747, + "learning_rate": 0.0004993476563080809, + "loss": 0.0012, + "step": 40230 + }, + { + "epoch": 60.51127819548872, + "grad_norm": 0.024139411747455597, + "learning_rate": 0.0004992897250651535, + "loss": 0.0011, + "step": 40240 + }, + { + "epoch": 60.526315789473685, + "grad_norm": 0.029365183785557747, + "learning_rate": 0.000499229333433282, + "loss": 0.0011, + "step": 40250 + }, + { + "epoch": 60.54135338345865, + "grad_norm": 0.02694946900010109, + "learning_rate": 0.0004991664820085074, + "loss": 0.0013, + "step": 40260 + }, + { + "epoch": 60.556390977443606, + "grad_norm": 0.01801641657948494, + "learning_rate": 0.000499101171411148, + "loss": 0.0011, + "step": 40270 + }, + { + "epoch": 60.57142857142857, + "grad_norm": 0.021255647763609886, + "learning_rate": 0.0004990334022857932, + "loss": 0.001, + "step": 40280 + }, + { + "epoch": 60.58646616541353, + "grad_norm": 0.020126348361372948, + "learning_rate": 0.0004989631753012964, + "loss": 0.0014, + "step": 40290 + }, + { + "epoch": 60.6015037593985, + "grad_norm": 0.01880335994064808, + "learning_rate": 0.00049889049115077, + "loss": 0.0012, + "step": 40300 + }, + { + "epoch": 60.61654135338346, + "grad_norm": 0.020995886996388435, + "learning_rate": 0.0004988153505515771, + "loss": 0.0011, + "step": 40310 + }, + { + "epoch": 60.63157894736842, + "grad_norm": 0.017119059339165688, + "learning_rate": 0.0004987377542453251, + "loss": 0.0016, + "step": 40320 + }, + { + "epoch": 60.64661654135338, + "grad_norm": 0.04048225283622742, + "learning_rate": 0.0004986577029978581, + "loss": 0.0012, + "step": 40330 + }, + { + "epoch": 60.661654135338345, + "grad_norm": 0.014520532451570034, + "learning_rate": 0.0004985751975992497, + "loss": 0.0012, + "step": 40340 + }, + { + "epoch": 60.67669172932331, + "grad_norm": 0.018836241215467453, + "learning_rate": 0.0004984902388637949, + "loss": 0.0013, + "step": 40350 + }, + { + "epoch": 60.69172932330827, + "grad_norm": 0.026123441755771637, + "learning_rate": 0.0004984028276300021, + "loss": 0.0013, + "step": 40360 + }, + { + "epoch": 60.70676691729323, + "grad_norm": 0.018893344327807426, + "learning_rate": 0.0004983129647605849, + "loss": 0.0013, + "step": 40370 + }, + { + "epoch": 60.721804511278194, + "grad_norm": 0.012534620240330696, + "learning_rate": 0.0004982206511424534, + "loss": 0.0011, + "step": 40380 + }, + { + "epoch": 60.73684210526316, + "grad_norm": 0.01753111369907856, + "learning_rate": 0.000498125887686706, + "loss": 0.001, + "step": 40390 + }, + { + "epoch": 60.75187969924812, + "grad_norm": 0.016191544011235237, + "learning_rate": 0.0004980286753286195, + "loss": 0.0012, + "step": 40400 + }, + { + "epoch": 60.766917293233085, + "grad_norm": 0.01594398356974125, + "learning_rate": 0.0004979290150276407, + "loss": 0.0012, + "step": 40410 + }, + { + "epoch": 60.78195488721804, + "grad_norm": 0.03318198397755623, + "learning_rate": 0.0004978269077673766, + "loss": 0.0013, + "step": 40420 + }, + { + "epoch": 60.796992481203006, + "grad_norm": 0.02540198341012001, + "learning_rate": 0.0004977223545555847, + "loss": 0.0013, + "step": 40430 + }, + { + "epoch": 60.81203007518797, + "grad_norm": 0.02130216732621193, + "learning_rate": 0.0004976153564241628, + "loss": 0.0015, + "step": 40440 + }, + { + "epoch": 60.82706766917293, + "grad_norm": 0.02429032512009144, + "learning_rate": 0.0004975059144291394, + "loss": 0.0013, + "step": 40450 + }, + { + "epoch": 60.8421052631579, + "grad_norm": 0.02775254286825657, + "learning_rate": 0.0004973940296506627, + "loss": 0.0011, + "step": 40460 + }, + { + "epoch": 60.857142857142854, + "grad_norm": 0.019704127684235573, + "learning_rate": 0.0004972797031929904, + "loss": 0.0011, + "step": 40470 + }, + { + "epoch": 60.87218045112782, + "grad_norm": 0.016386453062295914, + "learning_rate": 0.0004971629361844785, + "loss": 0.0012, + "step": 40480 + }, + { + "epoch": 60.88721804511278, + "grad_norm": 0.017072658985853195, + "learning_rate": 0.0004970437297775702, + "loss": 0.0012, + "step": 40490 + }, + { + "epoch": 60.902255639097746, + "grad_norm": 0.019210048019886017, + "learning_rate": 0.0004969220851487844, + "loss": 0.001, + "step": 40500 + }, + { + "epoch": 60.91729323308271, + "grad_norm": 0.019603153690695763, + "learning_rate": 0.0004967980034987048, + "loss": 0.0011, + "step": 40510 + }, + { + "epoch": 60.932330827067666, + "grad_norm": 0.02187393233180046, + "learning_rate": 0.000496671486051967, + "loss": 0.0011, + "step": 40520 + }, + { + "epoch": 60.94736842105263, + "grad_norm": 0.01793494075536728, + "learning_rate": 0.0004965425340572472, + "loss": 0.0012, + "step": 40530 + }, + { + "epoch": 60.962406015037594, + "grad_norm": 0.01727982610464096, + "learning_rate": 0.0004964111487872495, + "loss": 0.0014, + "step": 40540 + }, + { + "epoch": 60.97744360902256, + "grad_norm": 0.024438247084617615, + "learning_rate": 0.0004962773315386935, + "loss": 0.0008, + "step": 40550 + }, + { + "epoch": 60.99248120300752, + "grad_norm": 0.0200142003595829, + "learning_rate": 0.0004961410836323014, + "loss": 0.0013, + "step": 40560 + }, + { + "epoch": 61.00751879699248, + "grad_norm": 0.03610473871231079, + "learning_rate": 0.0004960024064127849, + "loss": 0.0012, + "step": 40570 + }, + { + "epoch": 61.02255639097744, + "grad_norm": 0.016493385657668114, + "learning_rate": 0.0004958613012488324, + "loss": 0.0009, + "step": 40580 + }, + { + "epoch": 61.037593984962406, + "grad_norm": 0.029424509033560753, + "learning_rate": 0.0004957177695330948, + "loss": 0.0011, + "step": 40590 + }, + { + "epoch": 61.05263157894737, + "grad_norm": 0.018477708101272583, + "learning_rate": 0.0004955718126821722, + "loss": 0.0011, + "step": 40600 + }, + { + "epoch": 61.067669172932334, + "grad_norm": 0.023900272324681282, + "learning_rate": 0.0004954234321365998, + "loss": 0.001, + "step": 40610 + }, + { + "epoch": 61.08270676691729, + "grad_norm": 0.019703850150108337, + "learning_rate": 0.0004952726293608335, + "loss": 0.0012, + "step": 40620 + }, + { + "epoch": 61.097744360902254, + "grad_norm": 0.016438821330666542, + "learning_rate": 0.0004951194058432361, + "loss": 0.0014, + "step": 40630 + }, + { + "epoch": 61.11278195488722, + "grad_norm": 0.020885130390524864, + "learning_rate": 0.0004949637630960618, + "loss": 0.0014, + "step": 40640 + }, + { + "epoch": 61.12781954887218, + "grad_norm": 0.01162684801965952, + "learning_rate": 0.0004948057026554415, + "loss": 0.0013, + "step": 40650 + }, + { + "epoch": 61.142857142857146, + "grad_norm": 0.019466811791062355, + "learning_rate": 0.000494645226081368, + "loss": 0.0009, + "step": 40660 + }, + { + "epoch": 61.1578947368421, + "grad_norm": 0.02157074771821499, + "learning_rate": 0.0004944823349576805, + "loss": 0.0013, + "step": 40670 + }, + { + "epoch": 61.17293233082707, + "grad_norm": 0.014531636610627174, + "learning_rate": 0.0004943170308920483, + "loss": 0.001, + "step": 40680 + }, + { + "epoch": 61.18796992481203, + "grad_norm": 0.01822495274245739, + "learning_rate": 0.0004941493155159562, + "loss": 0.0011, + "step": 40690 + }, + { + "epoch": 61.203007518796994, + "grad_norm": 0.030857494100928307, + "learning_rate": 0.0004939791904846869, + "loss": 0.0013, + "step": 40700 + }, + { + "epoch": 61.21804511278196, + "grad_norm": 0.01153195183724165, + "learning_rate": 0.0004938066574773058, + "loss": 0.001, + "step": 40710 + }, + { + "epoch": 61.233082706766915, + "grad_norm": 0.01682121679186821, + "learning_rate": 0.0004936317181966443, + "loss": 0.0011, + "step": 40720 + }, + { + "epoch": 61.24812030075188, + "grad_norm": 0.02484051138162613, + "learning_rate": 0.0004934543743692822, + "loss": 0.0011, + "step": 40730 + }, + { + "epoch": 61.26315789473684, + "grad_norm": 0.005320678930729628, + "learning_rate": 0.0004932746277455317, + "loss": 0.001, + "step": 40740 + }, + { + "epoch": 61.278195488721806, + "grad_norm": 0.026836758479475975, + "learning_rate": 0.0004930924800994192, + "loss": 0.0011, + "step": 40750 + }, + { + "epoch": 61.29323308270677, + "grad_norm": 0.02479587309062481, + "learning_rate": 0.0004929079332286685, + "loss": 0.0008, + "step": 40760 + }, + { + "epoch": 61.30827067669173, + "grad_norm": 0.029882650822401047, + "learning_rate": 0.0004927209889546828, + "loss": 0.0014, + "step": 40770 + }, + { + "epoch": 61.32330827067669, + "grad_norm": 0.02497199922800064, + "learning_rate": 0.0004925316491225265, + "loss": 0.0013, + "step": 40780 + }, + { + "epoch": 61.338345864661655, + "grad_norm": 0.020993946120142937, + "learning_rate": 0.0004923399156009073, + "loss": 0.0011, + "step": 40790 + }, + { + "epoch": 61.35338345864662, + "grad_norm": 0.01732916384935379, + "learning_rate": 0.0004921457902821578, + "loss": 0.0012, + "step": 40800 + }, + { + "epoch": 61.36842105263158, + "grad_norm": 0.0212508887052536, + "learning_rate": 0.0004919492750822163, + "loss": 0.0009, + "step": 40810 + }, + { + "epoch": 61.38345864661654, + "grad_norm": 0.019366322085261345, + "learning_rate": 0.0004917503719406087, + "loss": 0.0011, + "step": 40820 + }, + { + "epoch": 61.3984962406015, + "grad_norm": 0.017939582467079163, + "learning_rate": 0.0004915490828204287, + "loss": 0.0011, + "step": 40830 + }, + { + "epoch": 61.41353383458647, + "grad_norm": 0.014842836186289787, + "learning_rate": 0.0004913454097083185, + "loss": 0.0008, + "step": 40840 + }, + { + "epoch": 61.42857142857143, + "grad_norm": 0.019387437030673027, + "learning_rate": 0.0004911393546144495, + "loss": 0.001, + "step": 40850 + }, + { + "epoch": 61.443609022556394, + "grad_norm": 0.020788295194506645, + "learning_rate": 0.0004909309195725024, + "loss": 0.0013, + "step": 40860 + }, + { + "epoch": 61.45864661654135, + "grad_norm": 0.01120759453624487, + "learning_rate": 0.0004907201066396469, + "loss": 0.0012, + "step": 40870 + }, + { + "epoch": 61.473684210526315, + "grad_norm": 0.006767623592168093, + "learning_rate": 0.0004905069178965214, + "loss": 0.0012, + "step": 40880 + }, + { + "epoch": 61.48872180451128, + "grad_norm": 0.028576767072081566, + "learning_rate": 0.000490291355447213, + "loss": 0.0011, + "step": 40890 + }, + { + "epoch": 61.50375939849624, + "grad_norm": 0.020309116691350937, + "learning_rate": 0.0004900734214192358, + "loss": 0.0011, + "step": 40900 + }, + { + "epoch": 61.5187969924812, + "grad_norm": 0.024727851152420044, + "learning_rate": 0.0004898531179635108, + "loss": 0.0018, + "step": 40910 + }, + { + "epoch": 61.53383458646616, + "grad_norm": 0.010762731544673443, + "learning_rate": 0.0004896304472543439, + "loss": 0.0011, + "step": 40920 + }, + { + "epoch": 61.54887218045113, + "grad_norm": 0.023299219086766243, + "learning_rate": 0.0004894054114894055, + "loss": 0.0016, + "step": 40930 + }, + { + "epoch": 61.56390977443609, + "grad_norm": 0.02675299160182476, + "learning_rate": 0.0004891780128897077, + "loss": 0.0013, + "step": 40940 + }, + { + "epoch": 61.578947368421055, + "grad_norm": 0.015647241845726967, + "learning_rate": 0.0004889482536995825, + "loss": 0.0013, + "step": 40950 + }, + { + "epoch": 61.59398496240601, + "grad_norm": 0.021386684849858284, + "learning_rate": 0.0004887161361866607, + "loss": 0.0012, + "step": 40960 + }, + { + "epoch": 61.609022556390975, + "grad_norm": 0.014737218618392944, + "learning_rate": 0.0004884816626418484, + "loss": 0.0011, + "step": 40970 + }, + { + "epoch": 61.62406015037594, + "grad_norm": 0.011370296590030193, + "learning_rate": 0.0004882448353793048, + "loss": 0.0013, + "step": 40980 + }, + { + "epoch": 61.6390977443609, + "grad_norm": 0.012001187540590763, + "learning_rate": 0.00048800565673641917, + "loss": 0.001, + "step": 40990 + }, + { + "epoch": 61.65413533834587, + "grad_norm": 0.021175481379032135, + "learning_rate": 0.0004877641290737884, + "loss": 0.0013, + "step": 41000 + }, + { + "epoch": 61.65413533834587, + "eval_cer": 0.01909787157645337, + "eval_loss": 0.09734708815813065, + "eval_runtime": 158.0093, + "eval_samples_per_second": 101.431, + "eval_steps_per_second": 0.797, + "eval_wer": 0.06871487906181285, + "step": 41000 + }, + { + "epoch": 61.669172932330824, + "grad_norm": 0.015375157818198204, + "learning_rate": 0.0004875202547751929, + "loss": 0.0011, + "step": 41010 + }, + { + "epoch": 61.68421052631579, + "grad_norm": 0.03217661380767822, + "learning_rate": 0.0004872740362475737, + "loss": 0.0012, + "step": 41020 + }, + { + "epoch": 61.69924812030075, + "grad_norm": 0.017894305288791656, + "learning_rate": 0.000487025475921008, + "loss": 0.0013, + "step": 41030 + }, + { + "epoch": 61.714285714285715, + "grad_norm": 0.02510494366288185, + "learning_rate": 0.0004867745762486861, + "loss": 0.0011, + "step": 41040 + }, + { + "epoch": 61.72932330827068, + "grad_norm": 0.026588406413793564, + "learning_rate": 0.00048652133970688633, + "loss": 0.0013, + "step": 41050 + }, + { + "epoch": 61.744360902255636, + "grad_norm": 0.02553599514067173, + "learning_rate": 0.0004862657687949512, + "loss": 0.0012, + "step": 41060 + }, + { + "epoch": 61.7593984962406, + "grad_norm": 0.024215690791606903, + "learning_rate": 0.0004860078660352625, + "loss": 0.0014, + "step": 41070 + }, + { + "epoch": 61.774436090225564, + "grad_norm": 0.013548159971833229, + "learning_rate": 0.0004857476339732161, + "loss": 0.0013, + "step": 41080 + }, + { + "epoch": 61.78947368421053, + "grad_norm": 0.02118677832186222, + "learning_rate": 0.00048548507517719766, + "loss": 0.0015, + "step": 41090 + }, + { + "epoch": 61.80451127819549, + "grad_norm": 0.025293754413723946, + "learning_rate": 0.0004852201922385564, + "loss": 0.0014, + "step": 41100 + }, + { + "epoch": 61.81954887218045, + "grad_norm": 0.020893698558211327, + "learning_rate": 0.00048495298777157994, + "loss": 0.0011, + "step": 41110 + }, + { + "epoch": 61.83458646616541, + "grad_norm": 0.045009128749370575, + "learning_rate": 0.00048468346441346853, + "loss": 0.0016, + "step": 41120 + }, + { + "epoch": 61.849624060150376, + "grad_norm": 0.01843002624809742, + "learning_rate": 0.0004844116248243089, + "loss": 0.0011, + "step": 41130 + }, + { + "epoch": 61.86466165413534, + "grad_norm": 0.020083652809262276, + "learning_rate": 0.0004841374716870481, + "loss": 0.0018, + "step": 41140 + }, + { + "epoch": 61.8796992481203, + "grad_norm": 0.019440768286585808, + "learning_rate": 0.00048386100770746686, + "loss": 0.0013, + "step": 41150 + }, + { + "epoch": 61.89473684210526, + "grad_norm": 0.011379090137779713, + "learning_rate": 0.00048358223561415306, + "loss": 0.0011, + "step": 41160 + }, + { + "epoch": 61.909774436090224, + "grad_norm": 0.02967836521565914, + "learning_rate": 0.00048330115815847465, + "loss": 0.0012, + "step": 41170 + }, + { + "epoch": 61.92481203007519, + "grad_norm": 0.02160620503127575, + "learning_rate": 0.00048301777811455274, + "loss": 0.001, + "step": 41180 + }, + { + "epoch": 61.93984962406015, + "grad_norm": 0.022965043783187866, + "learning_rate": 0.0004827320982792339, + "loss": 0.0009, + "step": 41190 + }, + { + "epoch": 61.954887218045116, + "grad_norm": 0.015050956979393959, + "learning_rate": 0.00048244412147206283, + "loss": 0.0009, + "step": 41200 + }, + { + "epoch": 61.96992481203007, + "grad_norm": 0.01961139403283596, + "learning_rate": 0.00048215385053525434, + "loss": 0.0013, + "step": 41210 + }, + { + "epoch": 61.984962406015036, + "grad_norm": 0.030702726915478706, + "learning_rate": 0.00048186128833366536, + "loss": 0.0013, + "step": 41220 + }, + { + "epoch": 62.0, + "grad_norm": 0.011043844744563103, + "learning_rate": 0.0004815664377547667, + "loss": 0.0014, + "step": 41230 + }, + { + "epoch": 62.015037593984964, + "grad_norm": 0.022839678451418877, + "learning_rate": 0.0004812693017086145, + "loss": 0.0011, + "step": 41240 + }, + { + "epoch": 62.03007518796993, + "grad_norm": 0.025643352419137955, + "learning_rate": 0.0004809698831278217, + "loss": 0.0011, + "step": 41250 + }, + { + "epoch": 62.045112781954884, + "grad_norm": 0.021068887785077095, + "learning_rate": 0.0004806681849675287, + "loss": 0.0008, + "step": 41260 + }, + { + "epoch": 62.06015037593985, + "grad_norm": 0.028734799474477768, + "learning_rate": 0.00048036421020537464, + "loss": 0.0011, + "step": 41270 + }, + { + "epoch": 62.07518796992481, + "grad_norm": 0.010324062779545784, + "learning_rate": 0.0004800579618414676, + "loss": 0.0009, + "step": 41280 + }, + { + "epoch": 62.090225563909776, + "grad_norm": 0.015161421149969101, + "learning_rate": 0.0004797494428983553, + "loss": 0.0006, + "step": 41290 + }, + { + "epoch": 62.10526315789474, + "grad_norm": 0.04059359058737755, + "learning_rate": 0.00047943865642099525, + "loss": 0.0013, + "step": 41300 + }, + { + "epoch": 62.1203007518797, + "grad_norm": 0.016712263226509094, + "learning_rate": 0.00047912560547672453, + "loss": 0.001, + "step": 41310 + }, + { + "epoch": 62.13533834586466, + "grad_norm": 0.023055506870150566, + "learning_rate": 0.0004788102931552294, + "loss": 0.001, + "step": 41320 + }, + { + "epoch": 62.150375939849624, + "grad_norm": 0.013706817291676998, + "learning_rate": 0.0004784927225685153, + "loss": 0.0012, + "step": 41330 + }, + { + "epoch": 62.16541353383459, + "grad_norm": 0.022568322718143463, + "learning_rate": 0.00047817289685087575, + "loss": 0.001, + "step": 41340 + }, + { + "epoch": 62.18045112781955, + "grad_norm": 0.0382382869720459, + "learning_rate": 0.0004778508191588613, + "loss": 0.0011, + "step": 41350 + }, + { + "epoch": 62.19548872180451, + "grad_norm": 0.01500299759209156, + "learning_rate": 0.00047752649267124894, + "loss": 0.0008, + "step": 41360 + }, + { + "epoch": 62.21052631578947, + "grad_norm": 0.015839261934161186, + "learning_rate": 0.00047719992058901006, + "loss": 0.0012, + "step": 41370 + }, + { + "epoch": 62.225563909774436, + "grad_norm": 0.01836260035634041, + "learning_rate": 0.00047687110613527924, + "loss": 0.0009, + "step": 41380 + }, + { + "epoch": 62.2406015037594, + "grad_norm": 0.008675700053572655, + "learning_rate": 0.00047654005255532247, + "loss": 0.001, + "step": 41390 + }, + { + "epoch": 62.255639097744364, + "grad_norm": 0.023491863161325455, + "learning_rate": 0.0004762067631165049, + "loss": 0.001, + "step": 41400 + }, + { + "epoch": 62.27067669172932, + "grad_norm": 0.016619956120848656, + "learning_rate": 0.00047587124110825874, + "loss": 0.0011, + "step": 41410 + }, + { + "epoch": 62.285714285714285, + "grad_norm": 0.03215517848730087, + "learning_rate": 0.0004755334898420507, + "loss": 0.0013, + "step": 41420 + }, + { + "epoch": 62.30075187969925, + "grad_norm": 0.017982082441449165, + "learning_rate": 0.00047519351265134954, + "loss": 0.0009, + "step": 41430 + }, + { + "epoch": 62.31578947368421, + "grad_norm": 0.02117021009325981, + "learning_rate": 0.0004748513128915928, + "loss": 0.0012, + "step": 41440 + }, + { + "epoch": 62.330827067669176, + "grad_norm": 0.02954980544745922, + "learning_rate": 0.0004745068939401539, + "loss": 0.0012, + "step": 41450 + }, + { + "epoch": 62.34586466165413, + "grad_norm": 0.011857107281684875, + "learning_rate": 0.000474160259196309, + "loss": 0.0009, + "step": 41460 + }, + { + "epoch": 62.3609022556391, + "grad_norm": 0.019016103819012642, + "learning_rate": 0.0004738114120812029, + "loss": 0.001, + "step": 41470 + }, + { + "epoch": 62.37593984962406, + "grad_norm": 0.026660149917006493, + "learning_rate": 0.00047346035603781597, + "loss": 0.0012, + "step": 41480 + }, + { + "epoch": 62.390977443609025, + "grad_norm": 0.02270621806383133, + "learning_rate": 0.0004731070945309295, + "loss": 0.0009, + "step": 41490 + }, + { + "epoch": 62.40601503759399, + "grad_norm": 0.01408957690000534, + "learning_rate": 0.00047275163104709196, + "loss": 0.0011, + "step": 41500 + }, + { + "epoch": 62.421052631578945, + "grad_norm": 0.021228956058621407, + "learning_rate": 0.0004723939690945845, + "loss": 0.001, + "step": 41510 + }, + { + "epoch": 62.43609022556391, + "grad_norm": 0.023849118500947952, + "learning_rate": 0.00047203411220338615, + "loss": 0.0012, + "step": 41520 + }, + { + "epoch": 62.45112781954887, + "grad_norm": 0.028154712170362473, + "learning_rate": 0.0004716720639251392, + "loss": 0.0011, + "step": 41530 + }, + { + "epoch": 62.46616541353384, + "grad_norm": 0.02011668123304844, + "learning_rate": 0.0004713078278331138, + "loss": 0.001, + "step": 41540 + }, + { + "epoch": 62.4812030075188, + "grad_norm": 0.02323276363313198, + "learning_rate": 0.0004709414075221734, + "loss": 0.0014, + "step": 41550 + }, + { + "epoch": 62.49624060150376, + "grad_norm": 0.036698322743177414, + "learning_rate": 0.00047057280660873835, + "loss": 0.0013, + "step": 41560 + }, + { + "epoch": 62.51127819548872, + "grad_norm": 0.018733978271484375, + "learning_rate": 0.00047020202873075093, + "loss": 0.0014, + "step": 41570 + }, + { + "epoch": 62.526315789473685, + "grad_norm": 0.059648338705301285, + "learning_rate": 0.00046982907754763905, + "loss": 0.0011, + "step": 41580 + }, + { + "epoch": 62.54135338345865, + "grad_norm": 0.01812380738556385, + "learning_rate": 0.00046945395674028047, + "loss": 0.0013, + "step": 41590 + }, + { + "epoch": 62.556390977443606, + "grad_norm": 0.029180916026234627, + "learning_rate": 0.0004690766700109659, + "loss": 0.0011, + "step": 41600 + }, + { + "epoch": 62.57142857142857, + "grad_norm": 0.02123401314020157, + "learning_rate": 0.0004686972210833632, + "loss": 0.0011, + "step": 41610 + }, + { + "epoch": 62.58646616541353, + "grad_norm": 0.010789172723889351, + "learning_rate": 0.0004683156137024801, + "loss": 0.0008, + "step": 41620 + }, + { + "epoch": 62.6015037593985, + "grad_norm": 0.03166310489177704, + "learning_rate": 0.0004679318516346273, + "loss": 0.0011, + "step": 41630 + }, + { + "epoch": 62.61654135338346, + "grad_norm": 0.02352135255932808, + "learning_rate": 0.00046754593866738144, + "loss": 0.0012, + "step": 41640 + }, + { + "epoch": 62.63157894736842, + "grad_norm": 0.019083548337221146, + "learning_rate": 0.00046715787860954785, + "loss": 0.001, + "step": 41650 + }, + { + "epoch": 62.64661654135338, + "grad_norm": 0.013976410031318665, + "learning_rate": 0.00046676767529112254, + "loss": 0.0011, + "step": 41660 + }, + { + "epoch": 62.661654135338345, + "grad_norm": 0.018445804715156555, + "learning_rate": 0.00046637533256325476, + "loss": 0.0013, + "step": 41670 + }, + { + "epoch": 62.67669172932331, + "grad_norm": 0.031722575426101685, + "learning_rate": 0.0004659808542982088, + "loss": 0.0013, + "step": 41680 + }, + { + "epoch": 62.69172932330827, + "grad_norm": 0.024211544543504715, + "learning_rate": 0.000465584244389326, + "loss": 0.0009, + "step": 41690 + }, + { + "epoch": 62.70676691729323, + "grad_norm": 0.029264774173498154, + "learning_rate": 0.0004651855067509859, + "loss": 0.0008, + "step": 41700 + }, + { + "epoch": 62.721804511278194, + "grad_norm": 0.016588449478149414, + "learning_rate": 0.0004647846453185681, + "loss": 0.0014, + "step": 41710 + }, + { + "epoch": 62.73684210526316, + "grad_norm": 0.01970824785530567, + "learning_rate": 0.0004643816640484131, + "loss": 0.001, + "step": 41720 + }, + { + "epoch": 62.75187969924812, + "grad_norm": 0.062723308801651, + "learning_rate": 0.0004639765669177833, + "loss": 0.001, + "step": 41730 + }, + { + "epoch": 62.766917293233085, + "grad_norm": 0.023731106892228127, + "learning_rate": 0.0004635693579248238, + "loss": 0.0011, + "step": 41740 + }, + { + "epoch": 62.78195488721804, + "grad_norm": 0.022287411615252495, + "learning_rate": 0.00046316004108852305, + "loss": 0.0014, + "step": 41750 + }, + { + "epoch": 62.796992481203006, + "grad_norm": 0.016855228692293167, + "learning_rate": 0.000462748620448673, + "loss": 0.0012, + "step": 41760 + }, + { + "epoch": 62.81203007518797, + "grad_norm": 0.01900186575949192, + "learning_rate": 0.00046233510006582913, + "loss": 0.0012, + "step": 41770 + }, + { + "epoch": 62.82706766917293, + "grad_norm": 0.01961336098611355, + "learning_rate": 0.0004619194840212708, + "loss": 0.0012, + "step": 41780 + }, + { + "epoch": 62.8421052631579, + "grad_norm": 0.01943495310842991, + "learning_rate": 0.0004615017764169606, + "loss": 0.0011, + "step": 41790 + }, + { + "epoch": 62.857142857142854, + "grad_norm": 0.015791017562150955, + "learning_rate": 0.00046108198137550377, + "loss": 0.0014, + "step": 41800 + }, + { + "epoch": 62.87218045112782, + "grad_norm": 0.011287575587630272, + "learning_rate": 0.0004606601030401081, + "loss": 0.0006, + "step": 41810 + }, + { + "epoch": 62.88721804511278, + "grad_norm": 0.01793825626373291, + "learning_rate": 0.0004602361455745423, + "loss": 0.0011, + "step": 41820 + }, + { + "epoch": 62.902255639097746, + "grad_norm": 0.04121699556708336, + "learning_rate": 0.0004598101131630954, + "loss": 0.0013, + "step": 41830 + }, + { + "epoch": 62.91729323308271, + "grad_norm": 0.022918283939361572, + "learning_rate": 0.00045938201001053546, + "loss": 0.0011, + "step": 41840 + }, + { + "epoch": 62.932330827067666, + "grad_norm": 0.029571479186415672, + "learning_rate": 0.0004589518403420676, + "loss": 0.0012, + "step": 41850 + }, + { + "epoch": 62.94736842105263, + "grad_norm": 0.024929635226726532, + "learning_rate": 0.0004585196084032928, + "loss": 0.0012, + "step": 41860 + }, + { + "epoch": 62.962406015037594, + "grad_norm": 0.019344285130500793, + "learning_rate": 0.0004580853184601659, + "loss": 0.0012, + "step": 41870 + }, + { + "epoch": 62.97744360902256, + "grad_norm": 0.0122873205691576, + "learning_rate": 0.00045764897479895315, + "loss": 0.0012, + "step": 41880 + }, + { + "epoch": 62.99248120300752, + "grad_norm": 0.014207352884113789, + "learning_rate": 0.00045721058172619043, + "loss": 0.001, + "step": 41890 + }, + { + "epoch": 63.00751879699248, + "grad_norm": 0.014499494805932045, + "learning_rate": 0.00045677014356864043, + "loss": 0.0009, + "step": 41900 + }, + { + "epoch": 63.02255639097744, + "grad_norm": 0.025664687156677246, + "learning_rate": 0.00045632766467324995, + "loss": 0.001, + "step": 41910 + }, + { + "epoch": 63.037593984962406, + "grad_norm": 0.015042079612612724, + "learning_rate": 0.00045588314940710683, + "loss": 0.0008, + "step": 41920 + }, + { + "epoch": 63.05263157894737, + "grad_norm": 0.026022128760814667, + "learning_rate": 0.00045543660215739755, + "loss": 0.0009, + "step": 41930 + }, + { + "epoch": 63.067669172932334, + "grad_norm": 0.01704465039074421, + "learning_rate": 0.00045498802733136306, + "loss": 0.001, + "step": 41940 + }, + { + "epoch": 63.08270676691729, + "grad_norm": 0.021042119711637497, + "learning_rate": 0.0004545374293562559, + "loss": 0.0009, + "step": 41950 + }, + { + "epoch": 63.097744360902254, + "grad_norm": 0.025699840858578682, + "learning_rate": 0.00045408481267929604, + "loss": 0.001, + "step": 41960 + }, + { + "epoch": 63.11278195488722, + "grad_norm": 0.020087316632270813, + "learning_rate": 0.0004536301817676274, + "loss": 0.0009, + "step": 41970 + }, + { + "epoch": 63.12781954887218, + "grad_norm": 0.019194649532437325, + "learning_rate": 0.00045317354110827344, + "loss": 0.0011, + "step": 41980 + }, + { + "epoch": 63.142857142857146, + "grad_norm": 0.010901307687163353, + "learning_rate": 0.00045271489520809337, + "loss": 0.0009, + "step": 41990 + }, + { + "epoch": 63.1578947368421, + "grad_norm": 0.0123516283929348, + "learning_rate": 0.0004522542485937369, + "loss": 0.0009, + "step": 42000 + }, + { + "epoch": 63.1578947368421, + "eval_cer": 0.01851710804241886, + "eval_loss": 0.10246290266513824, + "eval_runtime": 162.6291, + "eval_samples_per_second": 98.549, + "eval_steps_per_second": 0.775, + "eval_wer": 0.0675731549174012, + "step": 42000 + }, + { + "epoch": 63.17293233082707, + "grad_norm": 0.03451775014400482, + "learning_rate": 0.00045179160581160005, + "loss": 0.0009, + "step": 42010 + }, + { + "epoch": 63.18796992481203, + "grad_norm": 0.02292013168334961, + "learning_rate": 0.00045132697142778044, + "loss": 0.0011, + "step": 42020 + }, + { + "epoch": 63.203007518796994, + "grad_norm": 0.009940207935869694, + "learning_rate": 0.0004508603500280319, + "loss": 0.0008, + "step": 42030 + }, + { + "epoch": 63.21804511278196, + "grad_norm": 0.016080431640148163, + "learning_rate": 0.00045039174621771915, + "loss": 0.0009, + "step": 42040 + }, + { + "epoch": 63.233082706766915, + "grad_norm": 0.01914866827428341, + "learning_rate": 0.0004499211646217727, + "loss": 0.0008, + "step": 42050 + }, + { + "epoch": 63.24812030075188, + "grad_norm": 0.020849157124757767, + "learning_rate": 0.00044944860988464276, + "loss": 0.001, + "step": 42060 + }, + { + "epoch": 63.26315789473684, + "grad_norm": 0.02834215760231018, + "learning_rate": 0.00044897408667025397, + "loss": 0.0008, + "step": 42070 + }, + { + "epoch": 63.278195488721806, + "grad_norm": 0.022441979497671127, + "learning_rate": 0.0004484975996619589, + "loss": 0.0009, + "step": 42080 + }, + { + "epoch": 63.29323308270677, + "grad_norm": 0.01562836579978466, + "learning_rate": 0.0004480191535624918, + "loss": 0.0007, + "step": 42090 + }, + { + "epoch": 63.30827067669173, + "grad_norm": 0.02243923209607601, + "learning_rate": 0.0004475387530939226, + "loss": 0.001, + "step": 42100 + }, + { + "epoch": 63.32330827067669, + "grad_norm": 0.020617099478840828, + "learning_rate": 0.00044705640299761004, + "loss": 0.0009, + "step": 42110 + }, + { + "epoch": 63.338345864661655, + "grad_norm": 0.02907959371805191, + "learning_rate": 0.0004465721080341547, + "loss": 0.0012, + "step": 42120 + }, + { + "epoch": 63.35338345864662, + "grad_norm": 0.02354019694030285, + "learning_rate": 0.0004460858729833525, + "loss": 0.0008, + "step": 42130 + }, + { + "epoch": 63.36842105263158, + "grad_norm": 0.034468431025743484, + "learning_rate": 0.000445597702644147, + "loss": 0.001, + "step": 42140 + }, + { + "epoch": 63.38345864661654, + "grad_norm": 0.026141628623008728, + "learning_rate": 0.0004451076018345824, + "loss": 0.0009, + "step": 42150 + }, + { + "epoch": 63.3984962406015, + "grad_norm": 0.014486228115856647, + "learning_rate": 0.00044461557539175587, + "loss": 0.0009, + "step": 42160 + }, + { + "epoch": 63.41353383458647, + "grad_norm": 0.023762382566928864, + "learning_rate": 0.00044412162817176966, + "loss": 0.0008, + "step": 42170 + }, + { + "epoch": 63.42857142857143, + "grad_norm": 0.032046277076005936, + "learning_rate": 0.00044362576504968344, + "loss": 0.0009, + "step": 42180 + }, + { + "epoch": 63.443609022556394, + "grad_norm": 0.014361979439854622, + "learning_rate": 0.0004431279909194661, + "loss": 0.0008, + "step": 42190 + }, + { + "epoch": 63.45864661654135, + "grad_norm": 0.01920711249113083, + "learning_rate": 0.0004426283106939473, + "loss": 0.0013, + "step": 42200 + }, + { + "epoch": 63.473684210526315, + "grad_norm": 0.01648193970322609, + "learning_rate": 0.0004421267293047692, + "loss": 0.0009, + "step": 42210 + }, + { + "epoch": 63.48872180451128, + "grad_norm": 0.024727199226617813, + "learning_rate": 0.00044162325170233745, + "loss": 0.001, + "step": 42220 + }, + { + "epoch": 63.50375939849624, + "grad_norm": 0.020732667297124863, + "learning_rate": 0.0004411178828557729, + "loss": 0.001, + "step": 42230 + }, + { + "epoch": 63.5187969924812, + "grad_norm": 0.01582776941359043, + "learning_rate": 0.000440610627752862, + "loss": 0.0009, + "step": 42240 + }, + { + "epoch": 63.53383458646616, + "grad_norm": 0.014358415268361568, + "learning_rate": 0.0004401014914000078, + "loss": 0.0011, + "step": 42250 + }, + { + "epoch": 63.54887218045113, + "grad_norm": 0.026920504868030548, + "learning_rate": 0.0004395904788221805, + "loss": 0.001, + "step": 42260 + }, + { + "epoch": 63.56390977443609, + "grad_norm": 0.018842779099941254, + "learning_rate": 0.00043907759506286797, + "loss": 0.001, + "step": 42270 + }, + { + "epoch": 63.578947368421055, + "grad_norm": 0.006627952214330435, + "learning_rate": 0.00043856284518402594, + "loss": 0.0009, + "step": 42280 + }, + { + "epoch": 63.59398496240601, + "grad_norm": 0.014269952662289143, + "learning_rate": 0.00043804623426602784, + "loss": 0.0008, + "step": 42290 + }, + { + "epoch": 63.609022556390975, + "grad_norm": 0.020527878776192665, + "learning_rate": 0.0004375277674076149, + "loss": 0.001, + "step": 42300 + }, + { + "epoch": 63.62406015037594, + "grad_norm": 0.013139592483639717, + "learning_rate": 0.0004370074497258456, + "loss": 0.0009, + "step": 42310 + }, + { + "epoch": 63.6390977443609, + "grad_norm": 0.019037745893001556, + "learning_rate": 0.00043648528635604556, + "loss": 0.0009, + "step": 42320 + }, + { + "epoch": 63.65413533834587, + "grad_norm": 0.019626647233963013, + "learning_rate": 0.0004359612824517563, + "loss": 0.0013, + "step": 42330 + }, + { + "epoch": 63.669172932330824, + "grad_norm": 0.030697088688611984, + "learning_rate": 0.0004354354431846848, + "loss": 0.001, + "step": 42340 + }, + { + "epoch": 63.68421052631579, + "grad_norm": 0.010423215106129646, + "learning_rate": 0.00043490777374465244, + "loss": 0.0008, + "step": 42350 + }, + { + "epoch": 63.69924812030075, + "grad_norm": 0.02007095143198967, + "learning_rate": 0.0004343782793395435, + "loss": 0.0009, + "step": 42360 + }, + { + "epoch": 63.714285714285715, + "grad_norm": 0.023491747677326202, + "learning_rate": 0.000433846965195254, + "loss": 0.0009, + "step": 42370 + }, + { + "epoch": 63.72932330827068, + "grad_norm": 0.013928968459367752, + "learning_rate": 0.00043331383655564003, + "loss": 0.0009, + "step": 42380 + }, + { + "epoch": 63.744360902255636, + "grad_norm": 0.007850827649235725, + "learning_rate": 0.00043277889868246605, + "loss": 0.0007, + "step": 42390 + }, + { + "epoch": 63.7593984962406, + "grad_norm": 0.023167263716459274, + "learning_rate": 0.00043224215685535287, + "loss": 0.0007, + "step": 42400 + }, + { + "epoch": 63.774436090225564, + "grad_norm": 0.011532480828464031, + "learning_rate": 0.0004317036163717257, + "loss": 0.0009, + "step": 42410 + }, + { + "epoch": 63.78947368421053, + "grad_norm": 0.00867086835205555, + "learning_rate": 0.0004311632825467617, + "loss": 0.0006, + "step": 42420 + }, + { + "epoch": 63.80451127819549, + "grad_norm": 0.013455228880047798, + "learning_rate": 0.00043062116071333745, + "loss": 0.0005, + "step": 42430 + }, + { + "epoch": 63.81954887218045, + "grad_norm": 0.023549769073724747, + "learning_rate": 0.00043007725622197675, + "loss": 0.0009, + "step": 42440 + }, + { + "epoch": 63.83458646616541, + "grad_norm": 0.014903437346220016, + "learning_rate": 0.0004295315744407972, + "loss": 0.0007, + "step": 42450 + }, + { + "epoch": 63.849624060150376, + "grad_norm": 0.01933128386735916, + "learning_rate": 0.0004289841207554578, + "loss": 0.001, + "step": 42460 + }, + { + "epoch": 63.86466165413534, + "grad_norm": 0.017243433743715286, + "learning_rate": 0.00042843490056910534, + "loss": 0.0008, + "step": 42470 + }, + { + "epoch": 63.8796992481203, + "grad_norm": 0.009420830756425858, + "learning_rate": 0.0004278839193023214, + "loss": 0.0011, + "step": 42480 + }, + { + "epoch": 63.89473684210526, + "grad_norm": 0.03037605620920658, + "learning_rate": 0.00042733118239306845, + "loss": 0.0008, + "step": 42490 + }, + { + "epoch": 63.909774436090224, + "grad_norm": 0.013440205715596676, + "learning_rate": 0.00042677669529663686, + "loss": 0.001, + "step": 42500 + }, + { + "epoch": 63.92481203007519, + "grad_norm": 0.0178923811763525, + "learning_rate": 0.00042622046348559034, + "loss": 0.0009, + "step": 42510 + }, + { + "epoch": 63.93984962406015, + "grad_norm": 0.026166247203946114, + "learning_rate": 0.00042566249244971235, + "loss": 0.0012, + "step": 42520 + }, + { + "epoch": 63.954887218045116, + "grad_norm": 0.020115451887249947, + "learning_rate": 0.0004251027876959516, + "loss": 0.0006, + "step": 42530 + }, + { + "epoch": 63.96992481203007, + "grad_norm": 0.016949446871876717, + "learning_rate": 0.00042454135474836817, + "loss": 0.0006, + "step": 42540 + }, + { + "epoch": 63.984962406015036, + "grad_norm": 0.013754605315625668, + "learning_rate": 0.00042397819914807855, + "loss": 0.0009, + "step": 42550 + }, + { + "epoch": 64.0, + "grad_norm": 0.02253049984574318, + "learning_rate": 0.00042341332645320126, + "loss": 0.001, + "step": 42560 + }, + { + "epoch": 64.01503759398496, + "grad_norm": 0.008425802923738956, + "learning_rate": 0.0004228467422388016, + "loss": 0.0007, + "step": 42570 + }, + { + "epoch": 64.03007518796993, + "grad_norm": 0.01965196244418621, + "learning_rate": 0.0004222784520968371, + "loss": 0.0008, + "step": 42580 + }, + { + "epoch": 64.04511278195488, + "grad_norm": 0.015570121817290783, + "learning_rate": 0.0004217084616361021, + "loss": 0.0007, + "step": 42590 + }, + { + "epoch": 64.06015037593986, + "grad_norm": 0.01732427254319191, + "learning_rate": 0.0004211367764821722, + "loss": 0.0009, + "step": 42600 + }, + { + "epoch": 64.07518796992481, + "grad_norm": 0.010992997325956821, + "learning_rate": 0.0004205634022773491, + "loss": 0.0006, + "step": 42610 + }, + { + "epoch": 64.09022556390977, + "grad_norm": 0.04898282513022423, + "learning_rate": 0.0004199883446806048, + "loss": 0.0011, + "step": 42620 + }, + { + "epoch": 64.10526315789474, + "grad_norm": 0.0177422147244215, + "learning_rate": 0.0004194116093675256, + "loss": 0.0008, + "step": 42630 + }, + { + "epoch": 64.1203007518797, + "grad_norm": 0.02133498154580593, + "learning_rate": 0.0004188332020302561, + "loss": 0.0007, + "step": 42640 + }, + { + "epoch": 64.13533834586467, + "grad_norm": 0.028325680643320084, + "learning_rate": 0.00041825312837744333, + "loss": 0.0008, + "step": 42650 + }, + { + "epoch": 64.15037593984962, + "grad_norm": 0.02480955608189106, + "learning_rate": 0.00041767139413418, + "loss": 0.0011, + "step": 42660 + }, + { + "epoch": 64.16541353383458, + "grad_norm": 0.021915679797530174, + "learning_rate": 0.0004170880050419483, + "loss": 0.0009, + "step": 42670 + }, + { + "epoch": 64.18045112781955, + "grad_norm": 0.02575453743338585, + "learning_rate": 0.0004165029668585629, + "loss": 0.0008, + "step": 42680 + }, + { + "epoch": 64.19548872180451, + "grad_norm": 0.01778576150536537, + "learning_rate": 0.00041591628535811464, + "loss": 0.0007, + "step": 42690 + }, + { + "epoch": 64.21052631578948, + "grad_norm": 0.023672569543123245, + "learning_rate": 0.00041532796633091297, + "loss": 0.0011, + "step": 42700 + }, + { + "epoch": 64.22556390977444, + "grad_norm": 0.014683867804706097, + "learning_rate": 0.0004147380155834293, + "loss": 0.0007, + "step": 42710 + }, + { + "epoch": 64.2406015037594, + "grad_norm": 0.02675030753016472, + "learning_rate": 0.0004141464389382391, + "loss": 0.0007, + "step": 42720 + }, + { + "epoch": 64.25563909774436, + "grad_norm": 0.013403667137026787, + "learning_rate": 0.0004135532422339653, + "loss": 0.0006, + "step": 42730 + }, + { + "epoch": 64.27067669172932, + "grad_norm": 0.011712467297911644, + "learning_rate": 0.00041295843132521973, + "loss": 0.0007, + "step": 42740 + }, + { + "epoch": 64.28571428571429, + "grad_norm": 0.024735508486628532, + "learning_rate": 0.0004123620120825459, + "loss": 0.0007, + "step": 42750 + }, + { + "epoch": 64.30075187969925, + "grad_norm": 0.015422341413795948, + "learning_rate": 0.0004117639903923611, + "loss": 0.0008, + "step": 42760 + }, + { + "epoch": 64.3157894736842, + "grad_norm": 0.017520209774374962, + "learning_rate": 0.00041116437215689785, + "loss": 0.0008, + "step": 42770 + }, + { + "epoch": 64.33082706766918, + "grad_norm": 0.024124018847942352, + "learning_rate": 0.00041056316329414613, + "loss": 0.0008, + "step": 42780 + }, + { + "epoch": 64.34586466165413, + "grad_norm": 0.02833697572350502, + "learning_rate": 0.0004099603697377946, + "loss": 0.0008, + "step": 42790 + }, + { + "epoch": 64.3609022556391, + "grad_norm": 0.021664608269929886, + "learning_rate": 0.00040935599743717243, + "loss": 0.0011, + "step": 42800 + }, + { + "epoch": 64.37593984962406, + "grad_norm": 0.014411985874176025, + "learning_rate": 0.0004087500523571902, + "loss": 0.0008, + "step": 42810 + }, + { + "epoch": 64.39097744360902, + "grad_norm": 0.024911068379878998, + "learning_rate": 0.0004081425404782811, + "loss": 0.0009, + "step": 42820 + }, + { + "epoch": 64.40601503759399, + "grad_norm": 0.011713022366166115, + "learning_rate": 0.0004075334677963423, + "loss": 0.0008, + "step": 42830 + }, + { + "epoch": 64.42105263157895, + "grad_norm": 0.016889724880456924, + "learning_rate": 0.00040692284032267515, + "loss": 0.0008, + "step": 42840 + }, + { + "epoch": 64.43609022556392, + "grad_norm": 0.010985138826072216, + "learning_rate": 0.00040631066408392636, + "loss": 0.0008, + "step": 42850 + }, + { + "epoch": 64.45112781954887, + "grad_norm": 0.01052536629140377, + "learning_rate": 0.00040569694512202815, + "loss": 0.0008, + "step": 42860 + }, + { + "epoch": 64.46616541353383, + "grad_norm": 0.01730651594698429, + "learning_rate": 0.00040508168949413904, + "loss": 0.0006, + "step": 42870 + }, + { + "epoch": 64.4812030075188, + "grad_norm": 0.020818671211600304, + "learning_rate": 0.0004044649032725836, + "loss": 0.001, + "step": 42880 + }, + { + "epoch": 64.49624060150376, + "grad_norm": 0.02159029059112072, + "learning_rate": 0.0004038465925447929, + "loss": 0.0008, + "step": 42890 + }, + { + "epoch": 64.51127819548873, + "grad_norm": 0.01798759214580059, + "learning_rate": 0.00040322676341324415, + "loss": 0.0006, + "step": 42900 + }, + { + "epoch": 64.52631578947368, + "grad_norm": 0.0425637811422348, + "learning_rate": 0.00040260542199540064, + "loss": 0.0009, + "step": 42910 + }, + { + "epoch": 64.54135338345864, + "grad_norm": 0.004783845506608486, + "learning_rate": 0.0004019825744236514, + "loss": 0.0005, + "step": 42920 + }, + { + "epoch": 64.55639097744361, + "grad_norm": 0.0076760598458349705, + "learning_rate": 0.00040135822684525036, + "loss": 0.0007, + "step": 42930 + }, + { + "epoch": 64.57142857142857, + "grad_norm": 0.01358152274042368, + "learning_rate": 0.00040073238542225623, + "loss": 0.0007, + "step": 42940 + }, + { + "epoch": 64.58646616541354, + "grad_norm": 0.010505531914532185, + "learning_rate": 0.00040010505633147106, + "loss": 0.0009, + "step": 42950 + }, + { + "epoch": 64.6015037593985, + "grad_norm": 0.014366254210472107, + "learning_rate": 0.0003994762457643797, + "loss": 0.0007, + "step": 42960 + }, + { + "epoch": 64.61654135338345, + "grad_norm": 0.025895560160279274, + "learning_rate": 0.00039884595992708877, + "loss": 0.001, + "step": 42970 + }, + { + "epoch": 64.63157894736842, + "grad_norm": 0.0162052009254694, + "learning_rate": 0.00039821420504026486, + "loss": 0.0007, + "step": 42980 + }, + { + "epoch": 64.64661654135338, + "grad_norm": 0.02845771610736847, + "learning_rate": 0.00039758098733907364, + "loss": 0.001, + "step": 42990 + }, + { + "epoch": 64.66165413533835, + "grad_norm": 0.019588768482208252, + "learning_rate": 0.0003969463130731183, + "loss": 0.0007, + "step": 43000 + }, + { + "epoch": 64.66165413533835, + "eval_cer": 0.01861367528109907, + "eval_loss": 0.10359231382608414, + "eval_runtime": 159.433, + "eval_samples_per_second": 100.525, + "eval_steps_per_second": 0.79, + "eval_wer": 0.06752617038470936, + "step": 43000 + }, + { + "epoch": 64.67669172932331, + "grad_norm": 0.021233852952718735, + "learning_rate": 0.0003963101885063776, + "loss": 0.0007, + "step": 43010 + }, + { + "epoch": 64.69172932330827, + "grad_norm": 0.014228448271751404, + "learning_rate": 0.00039567261991714406, + "loss": 0.0006, + "step": 43020 + }, + { + "epoch": 64.70676691729324, + "grad_norm": 0.023321596905589104, + "learning_rate": 0.00039503361359796235, + "loss": 0.0008, + "step": 43030 + }, + { + "epoch": 64.7218045112782, + "grad_norm": 0.005575632676482201, + "learning_rate": 0.0003943931758555669, + "loss": 0.0008, + "step": 43040 + }, + { + "epoch": 64.73684210526316, + "grad_norm": 0.024880660697817802, + "learning_rate": 0.0003937513130108197, + "loss": 0.0007, + "step": 43050 + }, + { + "epoch": 64.75187969924812, + "grad_norm": 0.011946773156523705, + "learning_rate": 0.00039310803139864777, + "loss": 0.0009, + "step": 43060 + }, + { + "epoch": 64.76691729323308, + "grad_norm": 0.009216031059622765, + "learning_rate": 0.00039246333736798095, + "loss": 0.0008, + "step": 43070 + }, + { + "epoch": 64.78195488721805, + "grad_norm": 0.018579736351966858, + "learning_rate": 0.0003918172372816892, + "loss": 0.0009, + "step": 43080 + }, + { + "epoch": 64.796992481203, + "grad_norm": 0.010375715792179108, + "learning_rate": 0.0003911697375165193, + "loss": 0.0006, + "step": 43090 + }, + { + "epoch": 64.81203007518798, + "grad_norm": 0.015251466073095798, + "learning_rate": 0.00039052084446303264, + "loss": 0.0008, + "step": 43100 + }, + { + "epoch": 64.82706766917293, + "grad_norm": 0.005869607906788588, + "learning_rate": 0.00038987056452554177, + "loss": 0.0006, + "step": 43110 + }, + { + "epoch": 64.84210526315789, + "grad_norm": 0.010210808366537094, + "learning_rate": 0.000389218904122047, + "loss": 0.0006, + "step": 43120 + }, + { + "epoch": 64.85714285714286, + "grad_norm": 0.010835377499461174, + "learning_rate": 0.00038856586968417353, + "loss": 0.001, + "step": 43130 + }, + { + "epoch": 64.87218045112782, + "grad_norm": 0.02107802778482437, + "learning_rate": 0.0003879114676571076, + "loss": 0.0008, + "step": 43140 + }, + { + "epoch": 64.88721804511279, + "grad_norm": 0.021816303953528404, + "learning_rate": 0.00038725570449953296, + "loss": 0.0008, + "step": 43150 + }, + { + "epoch": 64.90225563909775, + "grad_norm": 0.02859710156917572, + "learning_rate": 0.0003865985866835673, + "loss": 0.0009, + "step": 43160 + }, + { + "epoch": 64.9172932330827, + "grad_norm": 0.014979415573179722, + "learning_rate": 0.00038594012069469814, + "loss": 0.0005, + "step": 43170 + }, + { + "epoch": 64.93233082706767, + "grad_norm": 0.013493204489350319, + "learning_rate": 0.000385280313031719, + "loss": 0.0008, + "step": 43180 + }, + { + "epoch": 64.94736842105263, + "grad_norm": 0.012228677049279213, + "learning_rate": 0.00038461917020666506, + "loss": 0.0009, + "step": 43190 + }, + { + "epoch": 64.9624060150376, + "grad_norm": 0.009888396598398685, + "learning_rate": 0.00038395669874474915, + "loss": 0.0008, + "step": 43200 + }, + { + "epoch": 64.97744360902256, + "grad_norm": 0.014137768186628819, + "learning_rate": 0.0003832929051842972, + "loss": 0.001, + "step": 43210 + }, + { + "epoch": 64.99248120300751, + "grad_norm": 0.03024628572165966, + "learning_rate": 0.00038262779607668354, + "loss": 0.0008, + "step": 43220 + }, + { + "epoch": 65.00751879699249, + "grad_norm": 0.017753394320607185, + "learning_rate": 0.00038196137798626663, + "loss": 0.0007, + "step": 43230 + }, + { + "epoch": 65.02255639097744, + "grad_norm": 0.009066886268556118, + "learning_rate": 0.00038129365749032395, + "loss": 0.0006, + "step": 43240 + }, + { + "epoch": 65.0375939849624, + "grad_norm": 0.015957500785589218, + "learning_rate": 0.0003806246411789872, + "loss": 0.0006, + "step": 43250 + }, + { + "epoch": 65.05263157894737, + "grad_norm": 0.024859890341758728, + "learning_rate": 0.0003799543356551773, + "loss": 0.0007, + "step": 43260 + }, + { + "epoch": 65.06766917293233, + "grad_norm": 0.007380116730928421, + "learning_rate": 0.0003792827475345393, + "loss": 0.0006, + "step": 43270 + }, + { + "epoch": 65.0827067669173, + "grad_norm": 0.014953267760574818, + "learning_rate": 0.0003786098834453766, + "loss": 0.0007, + "step": 43280 + }, + { + "epoch": 65.09774436090225, + "grad_norm": 0.009193986654281616, + "learning_rate": 0.00037793575002858625, + "loss": 0.0004, + "step": 43290 + }, + { + "epoch": 65.11278195488721, + "grad_norm": 0.015467883087694645, + "learning_rate": 0.00037726035393759286, + "loss": 0.0007, + "step": 43300 + }, + { + "epoch": 65.12781954887218, + "grad_norm": 0.021067989990115166, + "learning_rate": 0.0003765837018382831, + "loss": 0.0007, + "step": 43310 + }, + { + "epoch": 65.14285714285714, + "grad_norm": 0.01462789997458458, + "learning_rate": 0.00037590580040894024, + "loss": 0.0005, + "step": 43320 + }, + { + "epoch": 65.15789473684211, + "grad_norm": 0.009192284196615219, + "learning_rate": 0.0003752266563401775, + "loss": 0.0006, + "step": 43330 + }, + { + "epoch": 65.17293233082707, + "grad_norm": 0.01765633560717106, + "learning_rate": 0.0003745462763348727, + "loss": 0.0008, + "step": 43340 + }, + { + "epoch": 65.18796992481202, + "grad_norm": 0.022234344854950905, + "learning_rate": 0.0003738646671081019, + "loss": 0.0008, + "step": 43350 + }, + { + "epoch": 65.203007518797, + "grad_norm": 0.01601152867078781, + "learning_rate": 0.0003731818353870729, + "loss": 0.0007, + "step": 43360 + }, + { + "epoch": 65.21804511278195, + "grad_norm": 0.013482065871357918, + "learning_rate": 0.00037249778791105916, + "loss": 0.0006, + "step": 43370 + }, + { + "epoch": 65.23308270676692, + "grad_norm": 0.020814530551433563, + "learning_rate": 0.0003718125314313331, + "loss": 0.0005, + "step": 43380 + }, + { + "epoch": 65.24812030075188, + "grad_norm": 0.020349469035863876, + "learning_rate": 0.0003711260727110995, + "loss": 0.0007, + "step": 43390 + }, + { + "epoch": 65.26315789473684, + "grad_norm": 0.018875813111662865, + "learning_rate": 0.0003704384185254288, + "loss": 0.0007, + "step": 43400 + }, + { + "epoch": 65.2781954887218, + "grad_norm": 0.0132759353145957, + "learning_rate": 0.00036974957566119027, + "loss": 0.0007, + "step": 43410 + }, + { + "epoch": 65.29323308270676, + "grad_norm": 0.005789619870483875, + "learning_rate": 0.0003690595509169848, + "loss": 0.0006, + "step": 43420 + }, + { + "epoch": 65.30827067669173, + "grad_norm": 0.02831275761127472, + "learning_rate": 0.00036836835110307803, + "loss": 0.0006, + "step": 43430 + }, + { + "epoch": 65.32330827067669, + "grad_norm": 0.016496188938617706, + "learning_rate": 0.0003676759830413332, + "loss": 0.0007, + "step": 43440 + }, + { + "epoch": 65.33834586466165, + "grad_norm": 0.02355986088514328, + "learning_rate": 0.00036698245356514336, + "loss": 0.0008, + "step": 43450 + }, + { + "epoch": 65.35338345864662, + "grad_norm": 0.01808469183743, + "learning_rate": 0.0003662877695193646, + "loss": 0.0006, + "step": 43460 + }, + { + "epoch": 65.36842105263158, + "grad_norm": 0.014524759724736214, + "learning_rate": 0.00036559193776024794, + "loss": 0.0007, + "step": 43470 + }, + { + "epoch": 65.38345864661655, + "grad_norm": 0.03294939175248146, + "learning_rate": 0.000364894965155372, + "loss": 0.0011, + "step": 43480 + }, + { + "epoch": 65.3984962406015, + "grad_norm": 0.02199738286435604, + "learning_rate": 0.00036419685858357485, + "loss": 0.0008, + "step": 43490 + }, + { + "epoch": 65.41353383458646, + "grad_norm": 0.037506211549043655, + "learning_rate": 0.00036349762493488667, + "loss": 0.0006, + "step": 43500 + }, + { + "epoch": 65.42857142857143, + "grad_norm": 0.03144306689500809, + "learning_rate": 0.00036279727111046127, + "loss": 0.0005, + "step": 43510 + }, + { + "epoch": 65.44360902255639, + "grad_norm": 0.02383388951420784, + "learning_rate": 0.0003620958040225081, + "loss": 0.0008, + "step": 43520 + }, + { + "epoch": 65.45864661654136, + "grad_norm": 0.0195087231695652, + "learning_rate": 0.0003613932305942241, + "loss": 0.0008, + "step": 43530 + }, + { + "epoch": 65.47368421052632, + "grad_norm": 0.015426869504153728, + "learning_rate": 0.0003606895577597254, + "loss": 0.0009, + "step": 43540 + }, + { + "epoch": 65.48872180451127, + "grad_norm": 0.009056270122528076, + "learning_rate": 0.0003599847924639788, + "loss": 0.0005, + "step": 43550 + }, + { + "epoch": 65.50375939849624, + "grad_norm": 0.01251070573925972, + "learning_rate": 0.00035927894166273323, + "loss": 0.0007, + "step": 43560 + }, + { + "epoch": 65.5187969924812, + "grad_norm": 0.010309994220733643, + "learning_rate": 0.0003585720123224512, + "loss": 0.0006, + "step": 43570 + }, + { + "epoch": 65.53383458646617, + "grad_norm": 0.017802013084292412, + "learning_rate": 0.00035786401142023975, + "loss": 0.0008, + "step": 43580 + }, + { + "epoch": 65.54887218045113, + "grad_norm": 0.010692677460610867, + "learning_rate": 0.0003571549459437821, + "loss": 0.0006, + "step": 43590 + }, + { + "epoch": 65.56390977443608, + "grad_norm": 0.015723105520009995, + "learning_rate": 0.0003564448228912682, + "loss": 0.0008, + "step": 43600 + }, + { + "epoch": 65.57894736842105, + "grad_norm": 0.007313680835068226, + "learning_rate": 0.0003557336492713258, + "loss": 0.0007, + "step": 43610 + }, + { + "epoch": 65.59398496240601, + "grad_norm": 0.019217679277062416, + "learning_rate": 0.00035502143210295163, + "loss": 0.0008, + "step": 43620 + }, + { + "epoch": 65.60902255639098, + "grad_norm": 0.014173259027302265, + "learning_rate": 0.0003543081784154414, + "loss": 0.0006, + "step": 43630 + }, + { + "epoch": 65.62406015037594, + "grad_norm": 0.014721768908202648, + "learning_rate": 0.0003535938952483211, + "loss": 0.0006, + "step": 43640 + }, + { + "epoch": 65.6390977443609, + "grad_norm": 0.014570224098861217, + "learning_rate": 0.00035287858965127723, + "loss": 0.0005, + "step": 43650 + }, + { + "epoch": 65.65413533834587, + "grad_norm": 0.014428780414164066, + "learning_rate": 0.0003521622686840873, + "loss": 0.0006, + "step": 43660 + }, + { + "epoch": 65.66917293233082, + "grad_norm": 0.015470949001610279, + "learning_rate": 0.00035144493941655, + "loss": 0.001, + "step": 43670 + }, + { + "epoch": 65.6842105263158, + "grad_norm": 0.010241092182695866, + "learning_rate": 0.00035072660892841566, + "loss": 0.0005, + "step": 43680 + }, + { + "epoch": 65.69924812030075, + "grad_norm": 0.015097121708095074, + "learning_rate": 0.00035000728430931616, + "loss": 0.0006, + "step": 43690 + }, + { + "epoch": 65.71428571428571, + "grad_norm": 0.016652211546897888, + "learning_rate": 0.00034928697265869515, + "loss": 0.0007, + "step": 43700 + }, + { + "epoch": 65.72932330827068, + "grad_norm": 0.018854904919862747, + "learning_rate": 0.0003485656810857378, + "loss": 0.0007, + "step": 43710 + }, + { + "epoch": 65.74436090225564, + "grad_norm": 0.02175341732800007, + "learning_rate": 0.00034784341670930066, + "loss": 0.0009, + "step": 43720 + }, + { + "epoch": 65.7593984962406, + "grad_norm": 0.017964890226721764, + "learning_rate": 0.00034712018665784155, + "loss": 0.001, + "step": 43730 + }, + { + "epoch": 65.77443609022556, + "grad_norm": 0.017855364829301834, + "learning_rate": 0.00034639599806934917, + "loss": 0.0009, + "step": 43740 + }, + { + "epoch": 65.78947368421052, + "grad_norm": 0.018917052075266838, + "learning_rate": 0.0003456708580912725, + "loss": 0.0005, + "step": 43750 + }, + { + "epoch": 65.80451127819549, + "grad_norm": 0.01624327525496483, + "learning_rate": 0.0003449447738804503, + "loss": 0.0008, + "step": 43760 + }, + { + "epoch": 65.81954887218045, + "grad_norm": 0.025193827226758003, + "learning_rate": 0.00034421775260304067, + "loss": 0.0006, + "step": 43770 + }, + { + "epoch": 65.83458646616542, + "grad_norm": 0.012459870427846909, + "learning_rate": 0.0003434898014344501, + "loss": 0.0006, + "step": 43780 + }, + { + "epoch": 65.84962406015038, + "grad_norm": 0.01215010229498148, + "learning_rate": 0.00034276092755926275, + "loss": 0.0006, + "step": 43790 + }, + { + "epoch": 65.86466165413533, + "grad_norm": 0.011906623840332031, + "learning_rate": 0.00034203113817116957, + "loss": 0.0005, + "step": 43800 + }, + { + "epoch": 65.8796992481203, + "grad_norm": 0.008557597175240517, + "learning_rate": 0.000341300440472897, + "loss": 0.0006, + "step": 43810 + }, + { + "epoch": 65.89473684210526, + "grad_norm": 0.017239412292838097, + "learning_rate": 0.0003405688416761364, + "loss": 0.0007, + "step": 43820 + }, + { + "epoch": 65.90977443609023, + "grad_norm": 0.01901654340326786, + "learning_rate": 0.0003398363490014727, + "loss": 0.0009, + "step": 43830 + }, + { + "epoch": 65.92481203007519, + "grad_norm": 0.020500924438238144, + "learning_rate": 0.00033910296967831267, + "loss": 0.0007, + "step": 43840 + }, + { + "epoch": 65.93984962406014, + "grad_norm": 0.026595190167427063, + "learning_rate": 0.00033836871094481433, + "loss": 0.0008, + "step": 43850 + }, + { + "epoch": 65.95488721804512, + "grad_norm": 0.010435215197503567, + "learning_rate": 0.00033763358004781474, + "loss": 0.0007, + "step": 43860 + }, + { + "epoch": 65.96992481203007, + "grad_norm": 0.01933489367365837, + "learning_rate": 0.0003368975842427592, + "loss": 0.0005, + "step": 43870 + }, + { + "epoch": 65.98496240601504, + "grad_norm": 0.012001696974039078, + "learning_rate": 0.00033616073079362923, + "loss": 0.0007, + "step": 43880 + }, + { + "epoch": 66.0, + "grad_norm": 0.03030433878302574, + "learning_rate": 0.0003354230269728709, + "loss": 0.0007, + "step": 43890 + }, + { + "epoch": 66.01503759398496, + "grad_norm": 0.016028081998229027, + "learning_rate": 0.0003346844800613229, + "loss": 0.0006, + "step": 43900 + }, + { + "epoch": 66.03007518796993, + "grad_norm": 0.0164579339325428, + "learning_rate": 0.00033394509734814516, + "loss": 0.0007, + "step": 43910 + }, + { + "epoch": 66.04511278195488, + "grad_norm": 0.013365254737436771, + "learning_rate": 0.00033320488613074666, + "loss": 0.0007, + "step": 43920 + }, + { + "epoch": 66.06015037593986, + "grad_norm": 0.018845556303858757, + "learning_rate": 0.0003324638537147132, + "loss": 0.0006, + "step": 43930 + }, + { + "epoch": 66.07518796992481, + "grad_norm": 0.007265174761414528, + "learning_rate": 0.0003317220074137356, + "loss": 0.0006, + "step": 43940 + }, + { + "epoch": 66.09022556390977, + "grad_norm": 0.07480914890766144, + "learning_rate": 0.00033097935454953737, + "loss": 0.0007, + "step": 43950 + }, + { + "epoch": 66.10526315789474, + "grad_norm": 0.023977505043148994, + "learning_rate": 0.00033023590245180237, + "loss": 0.0006, + "step": 43960 + }, + { + "epoch": 66.1203007518797, + "grad_norm": 0.010117200203239918, + "learning_rate": 0.0003294916584581027, + "loss": 0.0005, + "step": 43970 + }, + { + "epoch": 66.13533834586467, + "grad_norm": 0.03143252432346344, + "learning_rate": 0.0003287466299138262, + "loss": 0.0008, + "step": 43980 + }, + { + "epoch": 66.15037593984962, + "grad_norm": 0.025035768747329712, + "learning_rate": 0.0003280008241721038, + "loss": 0.0006, + "step": 43990 + }, + { + "epoch": 66.16541353383458, + "grad_norm": 0.01127985306084156, + "learning_rate": 0.00032725424859373687, + "loss": 0.0007, + "step": 44000 + }, + { + "epoch": 66.16541353383458, + "eval_cer": 0.01810635725253965, + "eval_loss": 0.10616657882928848, + "eval_runtime": 160.2522, + "eval_samples_per_second": 100.011, + "eval_steps_per_second": 0.786, + "eval_wer": 0.06692006991298464, + "step": 44000 + }, + { + "epoch": 66.18045112781955, + "grad_norm": 0.012798226438462734, + "learning_rate": 0.00032650691054712523, + "loss": 0.0006, + "step": 44010 + }, + { + "epoch": 66.19548872180451, + "grad_norm": 0.018899250775575638, + "learning_rate": 0.00032575881740819353, + "loss": 0.0005, + "step": 44020 + }, + { + "epoch": 66.21052631578948, + "grad_norm": 0.00531221367418766, + "learning_rate": 0.00032500997656031907, + "loss": 0.0005, + "step": 44030 + }, + { + "epoch": 66.22556390977444, + "grad_norm": 0.02031558007001877, + "learning_rate": 0.0003242603953942587, + "loss": 0.0007, + "step": 44040 + }, + { + "epoch": 66.2406015037594, + "grad_norm": 0.010291735641658306, + "learning_rate": 0.000323510081308076, + "loss": 0.0005, + "step": 44050 + }, + { + "epoch": 66.25563909774436, + "grad_norm": 0.021605392917990685, + "learning_rate": 0.0003227590417070679, + "loss": 0.0007, + "step": 44060 + }, + { + "epoch": 66.27067669172932, + "grad_norm": 0.01669122651219368, + "learning_rate": 0.00032200728400369233, + "loss": 0.0008, + "step": 44070 + }, + { + "epoch": 66.28571428571429, + "grad_norm": 0.012600567191839218, + "learning_rate": 0.00032125481561749405, + "loss": 0.0005, + "step": 44080 + }, + { + "epoch": 66.30075187969925, + "grad_norm": 0.005582880694419146, + "learning_rate": 0.0003205016439750323, + "loss": 0.0005, + "step": 44090 + }, + { + "epoch": 66.3157894736842, + "grad_norm": 0.018399016931653023, + "learning_rate": 0.00031974777650980735, + "loss": 0.0007, + "step": 44100 + }, + { + "epoch": 66.33082706766918, + "grad_norm": 0.008783812634646893, + "learning_rate": 0.0003189932206621865, + "loss": 0.0005, + "step": 44110 + }, + { + "epoch": 66.34586466165413, + "grad_norm": 0.007357433903962374, + "learning_rate": 0.00031823798387933133, + "loss": 0.0006, + "step": 44120 + }, + { + "epoch": 66.3609022556391, + "grad_norm": 0.01924520544707775, + "learning_rate": 0.00031748207361512415, + "loss": 0.0006, + "step": 44130 + }, + { + "epoch": 66.37593984962406, + "grad_norm": 0.023862361907958984, + "learning_rate": 0.00031672549733009395, + "loss": 0.0006, + "step": 44140 + }, + { + "epoch": 66.39097744360902, + "grad_norm": 0.0201218593865633, + "learning_rate": 0.00031596826249134324, + "loss": 0.0008, + "step": 44150 + }, + { + "epoch": 66.40601503759399, + "grad_norm": 0.021591460332274437, + "learning_rate": 0.0003152103765724743, + "loss": 0.0007, + "step": 44160 + }, + { + "epoch": 66.42105263157895, + "grad_norm": 0.01150228176265955, + "learning_rate": 0.000314451847053515, + "loss": 0.0004, + "step": 44170 + }, + { + "epoch": 66.43609022556392, + "grad_norm": 0.021153416484594345, + "learning_rate": 0.00031369268142084555, + "loss": 0.0006, + "step": 44180 + }, + { + "epoch": 66.45112781954887, + "grad_norm": 0.01422085054218769, + "learning_rate": 0.0003129328871671243, + "loss": 0.0006, + "step": 44190 + }, + { + "epoch": 66.46616541353383, + "grad_norm": 0.01687667891383171, + "learning_rate": 0.0003121724717912138, + "loss": 0.0006, + "step": 44200 + }, + { + "epoch": 66.4812030075188, + "grad_norm": 0.020219463855028152, + "learning_rate": 0.0003114114427981066, + "loss": 0.0005, + "step": 44210 + }, + { + "epoch": 66.49624060150376, + "grad_norm": 0.016208436340093613, + "learning_rate": 0.0003106498076988519, + "loss": 0.0006, + "step": 44220 + }, + { + "epoch": 66.51127819548873, + "grad_norm": 0.015383698046207428, + "learning_rate": 0.0003098875740104805, + "loss": 0.0005, + "step": 44230 + }, + { + "epoch": 66.52631578947368, + "grad_norm": 0.015269642695784569, + "learning_rate": 0.0003091247492559312, + "loss": 0.0005, + "step": 44240 + }, + { + "epoch": 66.54135338345864, + "grad_norm": 0.016474300995469093, + "learning_rate": 0.0003083613409639764, + "loss": 0.0006, + "step": 44250 + }, + { + "epoch": 66.55639097744361, + "grad_norm": 0.013877849094569683, + "learning_rate": 0.00030759735666914767, + "loss": 0.0006, + "step": 44260 + }, + { + "epoch": 66.57142857142857, + "grad_norm": 0.01549141202121973, + "learning_rate": 0.0003068328039116616, + "loss": 0.0007, + "step": 44270 + }, + { + "epoch": 66.58646616541354, + "grad_norm": 0.01649453118443489, + "learning_rate": 0.00030606769023734534, + "loss": 0.0006, + "step": 44280 + }, + { + "epoch": 66.6015037593985, + "grad_norm": 0.022783882915973663, + "learning_rate": 0.00030530202319756184, + "loss": 0.0005, + "step": 44290 + }, + { + "epoch": 66.61654135338345, + "grad_norm": 0.014149666763842106, + "learning_rate": 0.0003045358103491357, + "loss": 0.0008, + "step": 44300 + }, + { + "epoch": 66.63157894736842, + "grad_norm": 0.017091304063796997, + "learning_rate": 0.0003037690592542784, + "loss": 0.0006, + "step": 44310 + }, + { + "epoch": 66.64661654135338, + "grad_norm": 0.01875300332903862, + "learning_rate": 0.00030300177748051373, + "loss": 0.0007, + "step": 44320 + }, + { + "epoch": 66.66165413533835, + "grad_norm": 0.018541693687438965, + "learning_rate": 0.0003022339726006029, + "loss": 0.0004, + "step": 44330 + }, + { + "epoch": 66.67669172932331, + "grad_norm": 0.021340930834412575, + "learning_rate": 0.00030146565219247033, + "loss": 0.0005, + "step": 44340 + }, + { + "epoch": 66.69172932330827, + "grad_norm": 0.02437593601644039, + "learning_rate": 0.0003006968238391281, + "loss": 0.0007, + "step": 44350 + }, + { + "epoch": 66.70676691729324, + "grad_norm": 0.011975240893661976, + "learning_rate": 0.0002999274951286017, + "loss": 0.0004, + "step": 44360 + }, + { + "epoch": 66.7218045112782, + "grad_norm": 0.009479226544499397, + "learning_rate": 0.000299157673653855, + "loss": 0.0005, + "step": 44370 + }, + { + "epoch": 66.73684210526316, + "grad_norm": 0.010109545662999153, + "learning_rate": 0.00029838736701271514, + "loss": 0.0007, + "step": 44380 + }, + { + "epoch": 66.75187969924812, + "grad_norm": 0.0035245949402451515, + "learning_rate": 0.0002976165828077975, + "loss": 0.0005, + "step": 44390 + }, + { + "epoch": 66.76691729323308, + "grad_norm": 0.012983671389520168, + "learning_rate": 0.0002968453286464312, + "loss": 0.0004, + "step": 44400 + }, + { + "epoch": 66.78195488721805, + "grad_norm": 0.006246019620448351, + "learning_rate": 0.0002960736121405834, + "loss": 0.0006, + "step": 44410 + }, + { + "epoch": 66.796992481203, + "grad_norm": 0.009627862833440304, + "learning_rate": 0.0002953014409067844, + "loss": 0.0004, + "step": 44420 + }, + { + "epoch": 66.81203007518798, + "grad_norm": 0.01761075295507908, + "learning_rate": 0.0002945288225660525, + "loss": 0.0007, + "step": 44430 + }, + { + "epoch": 66.82706766917293, + "grad_norm": 0.010367386974394321, + "learning_rate": 0.00029375576474381903, + "loss": 0.0009, + "step": 44440 + }, + { + "epoch": 66.84210526315789, + "grad_norm": 0.0070490664802491665, + "learning_rate": 0.0002929822750698524, + "loss": 0.0005, + "step": 44450 + }, + { + "epoch": 66.85714285714286, + "grad_norm": 0.008255287073552608, + "learning_rate": 0.00029220836117818346, + "loss": 0.0008, + "step": 44460 + }, + { + "epoch": 66.87218045112782, + "grad_norm": 0.005265123210847378, + "learning_rate": 0.00029143403070702994, + "loss": 0.0007, + "step": 44470 + }, + { + "epoch": 66.88721804511279, + "grad_norm": 0.029706638306379318, + "learning_rate": 0.00029065929129872095, + "loss": 0.0007, + "step": 44480 + }, + { + "epoch": 66.90225563909775, + "grad_norm": 0.01875491626560688, + "learning_rate": 0.0002898841505996216, + "loss": 0.0005, + "step": 44490 + }, + { + "epoch": 66.9172932330827, + "grad_norm": 0.004413621034473181, + "learning_rate": 0.00028910861626005774, + "loss": 0.0006, + "step": 44500 + }, + { + "epoch": 66.93233082706767, + "grad_norm": 0.02159472368657589, + "learning_rate": 0.00028833269593424017, + "loss": 0.0006, + "step": 44510 + }, + { + "epoch": 66.94736842105263, + "grad_norm": 0.012431158684194088, + "learning_rate": 0.0002875563972801893, + "loss": 0.0006, + "step": 44520 + }, + { + "epoch": 66.9624060150376, + "grad_norm": 0.02110169269144535, + "learning_rate": 0.0002867797279596593, + "loss": 0.0009, + "step": 44530 + }, + { + "epoch": 66.97744360902256, + "grad_norm": 0.014925649389624596, + "learning_rate": 0.00028600269563806304, + "loss": 0.0007, + "step": 44540 + }, + { + "epoch": 66.99248120300751, + "grad_norm": 0.008735325187444687, + "learning_rate": 0.00028522530798439564, + "loss": 0.0004, + "step": 44550 + }, + { + "epoch": 67.00751879699249, + "grad_norm": 0.014745322987437248, + "learning_rate": 0.0002844475726711595, + "loss": 0.0005, + "step": 44560 + }, + { + "epoch": 67.02255639097744, + "grad_norm": 0.012826160527765751, + "learning_rate": 0.00028366949737428814, + "loss": 0.0005, + "step": 44570 + }, + { + "epoch": 67.0375939849624, + "grad_norm": 0.01323269959539175, + "learning_rate": 0.00028289108977307066, + "loss": 0.0004, + "step": 44580 + }, + { + "epoch": 67.05263157894737, + "grad_norm": 0.013878796249628067, + "learning_rate": 0.00028211235755007575, + "loss": 0.0004, + "step": 44590 + }, + { + "epoch": 67.06766917293233, + "grad_norm": 0.010835380293428898, + "learning_rate": 0.00028133330839107606, + "loss": 0.0004, + "step": 44600 + }, + { + "epoch": 67.0827067669173, + "grad_norm": 0.011065471917390823, + "learning_rate": 0.00028055394998497237, + "loss": 0.0005, + "step": 44610 + }, + { + "epoch": 67.09774436090225, + "grad_norm": 0.014592123217880726, + "learning_rate": 0.00027977429002371744, + "loss": 0.0005, + "step": 44620 + }, + { + "epoch": 67.11278195488721, + "grad_norm": 0.0045190732926130295, + "learning_rate": 0.00027899433620224033, + "loss": 0.0004, + "step": 44630 + }, + { + "epoch": 67.12781954887218, + "grad_norm": 0.016830384731292725, + "learning_rate": 0.0002782140962183704, + "loss": 0.0005, + "step": 44640 + }, + { + "epoch": 67.14285714285714, + "grad_norm": 0.010263838805258274, + "learning_rate": 0.0002774335777727613, + "loss": 0.0006, + "step": 44650 + }, + { + "epoch": 67.15789473684211, + "grad_norm": 0.0034704934805631638, + "learning_rate": 0.00027665278856881496, + "loss": 0.0005, + "step": 44660 + }, + { + "epoch": 67.17293233082707, + "grad_norm": 0.013030853122472763, + "learning_rate": 0.00027587173631260563, + "loss": 0.0006, + "step": 44670 + }, + { + "epoch": 67.18796992481202, + "grad_norm": 0.0029409038834273815, + "learning_rate": 0.0002750904287128037, + "loss": 0.0005, + "step": 44680 + }, + { + "epoch": 67.203007518797, + "grad_norm": 0.009565877728164196, + "learning_rate": 0.00027430887348059993, + "loss": 0.0004, + "step": 44690 + }, + { + "epoch": 67.21804511278195, + "grad_norm": 0.01395465712994337, + "learning_rate": 0.0002735270783296286, + "loss": 0.0004, + "step": 44700 + }, + { + "epoch": 67.23308270676692, + "grad_norm": 0.0142797427251935, + "learning_rate": 0.0002727450509758925, + "loss": 0.0004, + "step": 44710 + }, + { + "epoch": 67.24812030075188, + "grad_norm": 0.005255323369055986, + "learning_rate": 0.00027196279913768587, + "loss": 0.0003, + "step": 44720 + }, + { + "epoch": 67.26315789473684, + "grad_norm": 0.007272629998624325, + "learning_rate": 0.0002711803305355184, + "loss": 0.0005, + "step": 44730 + }, + { + "epoch": 67.2781954887218, + "grad_norm": 0.010265227407217026, + "learning_rate": 0.00027039765289203944, + "loss": 0.0004, + "step": 44740 + }, + { + "epoch": 67.29323308270676, + "grad_norm": 0.014171719551086426, + "learning_rate": 0.00026961477393196127, + "loss": 0.0005, + "step": 44750 + }, + { + "epoch": 67.30827067669173, + "grad_norm": 0.01433984562754631, + "learning_rate": 0.00026883170138198323, + "loss": 0.0004, + "step": 44760 + }, + { + "epoch": 67.32330827067669, + "grad_norm": 0.014160205610096455, + "learning_rate": 0.00026804844297071524, + "loss": 0.0007, + "step": 44770 + }, + { + "epoch": 67.33834586466165, + "grad_norm": 0.012176107615232468, + "learning_rate": 0.0002672650064286015, + "loss": 0.0005, + "step": 44780 + }, + { + "epoch": 67.35338345864662, + "grad_norm": 0.011361782439053059, + "learning_rate": 0.0002664813994878445, + "loss": 0.0004, + "step": 44790 + }, + { + "epoch": 67.36842105263158, + "grad_norm": 0.008113673888146877, + "learning_rate": 0.0002656976298823284, + "loss": 0.0004, + "step": 44800 + }, + { + "epoch": 67.38345864661655, + "grad_norm": 0.009716987609863281, + "learning_rate": 0.0002649137053475427, + "loss": 0.0005, + "step": 44810 + }, + { + "epoch": 67.3984962406015, + "grad_norm": 0.011081482283771038, + "learning_rate": 0.0002641296336205062, + "loss": 0.0005, + "step": 44820 + }, + { + "epoch": 67.41353383458646, + "grad_norm": 0.012672477401793003, + "learning_rate": 0.0002633454224396901, + "loss": 0.0004, + "step": 44830 + }, + { + "epoch": 67.42857142857143, + "grad_norm": 0.013974172063171864, + "learning_rate": 0.0002625610795449424, + "loss": 0.0005, + "step": 44840 + }, + { + "epoch": 67.44360902255639, + "grad_norm": 0.009405584074556828, + "learning_rate": 0.00026177661267741067, + "loss": 0.0004, + "step": 44850 + }, + { + "epoch": 67.45864661654136, + "grad_norm": 0.004881350789219141, + "learning_rate": 0.0002609920295794662, + "loss": 0.0005, + "step": 44860 + }, + { + "epoch": 67.47368421052632, + "grad_norm": 0.013648021966218948, + "learning_rate": 0.00026020733799462755, + "loss": 0.0005, + "step": 44870 + }, + { + "epoch": 67.48872180451127, + "grad_norm": 0.00875017512589693, + "learning_rate": 0.0002594225456674837, + "loss": 0.0005, + "step": 44880 + }, + { + "epoch": 67.50375939849624, + "grad_norm": 0.026211164891719818, + "learning_rate": 0.00025863766034361815, + "loss": 0.0006, + "step": 44890 + }, + { + "epoch": 67.5187969924812, + "grad_norm": 0.011105847544968128, + "learning_rate": 0.00025785268976953206, + "loss": 0.0005, + "step": 44900 + }, + { + "epoch": 67.53383458646617, + "grad_norm": 0.013042529113590717, + "learning_rate": 0.00025706764169256837, + "loss": 0.0005, + "step": 44910 + }, + { + "epoch": 67.54887218045113, + "grad_norm": 0.006796063855290413, + "learning_rate": 0.0002562825238608344, + "loss": 0.0005, + "step": 44920 + }, + { + "epoch": 67.56390977443608, + "grad_norm": 0.014797746203839779, + "learning_rate": 0.0002554973440231263, + "loss": 0.0006, + "step": 44930 + }, + { + "epoch": 67.57894736842105, + "grad_norm": 0.014406909234821796, + "learning_rate": 0.0002547121099288521, + "loss": 0.0005, + "step": 44940 + }, + { + "epoch": 67.59398496240601, + "grad_norm": 0.020130963996052742, + "learning_rate": 0.0002539268293279552, + "loss": 0.0005, + "step": 44950 + }, + { + "epoch": 67.60902255639098, + "grad_norm": 0.00767687289044261, + "learning_rate": 0.0002531415099708382, + "loss": 0.0002, + "step": 44960 + }, + { + "epoch": 67.62406015037594, + "grad_norm": 0.014985785819590092, + "learning_rate": 0.0002523561596082861, + "loss": 0.0004, + "step": 44970 + }, + { + "epoch": 67.6390977443609, + "grad_norm": 0.013695353642106056, + "learning_rate": 0.00025157078599138976, + "loss": 0.0004, + "step": 44980 + }, + { + "epoch": 67.65413533834587, + "grad_norm": 0.01647520437836647, + "learning_rate": 0.0002507853968714699, + "loss": 0.0007, + "step": 44990 + }, + { + "epoch": 67.66917293233082, + "grad_norm": 0.023190142586827278, + "learning_rate": 0.00025, + "loss": 0.0004, + "step": 45000 + }, + { + "epoch": 67.66917293233082, + "eval_cer": 0.01820700479707959, + "eval_loss": 0.10867446660995483, + "eval_runtime": 162.0536, + "eval_samples_per_second": 98.899, + "eval_steps_per_second": 0.778, + "eval_wer": 0.06688718074010036, + "step": 45000 + }, + { + "epoch": 67.6842105263158, + "grad_norm": 0.012358726933598518, + "learning_rate": 0.0002492146031285301, + "loss": 0.0004, + "step": 45010 + }, + { + "epoch": 67.69924812030075, + "grad_norm": 0.012489447370171547, + "learning_rate": 0.00024842921400861025, + "loss": 0.0004, + "step": 45020 + }, + { + "epoch": 67.71428571428571, + "grad_norm": 0.007779994979500771, + "learning_rate": 0.000247643840391714, + "loss": 0.0004, + "step": 45030 + }, + { + "epoch": 67.72932330827068, + "grad_norm": 0.008729352615773678, + "learning_rate": 0.0002468584900291618, + "loss": 0.0004, + "step": 45040 + }, + { + "epoch": 67.74436090225564, + "grad_norm": 0.007784692104905844, + "learning_rate": 0.0002460731706720449, + "loss": 0.0004, + "step": 45050 + }, + { + "epoch": 67.7593984962406, + "grad_norm": 0.012112563475966454, + "learning_rate": 0.000245287890071148, + "loss": 0.0004, + "step": 45060 + }, + { + "epoch": 67.77443609022556, + "grad_norm": 0.012248797342181206, + "learning_rate": 0.00024450265597687374, + "loss": 0.0004, + "step": 45070 + }, + { + "epoch": 67.78947368421052, + "grad_norm": 0.008206418715417385, + "learning_rate": 0.00024371747613916565, + "loss": 0.0004, + "step": 45080 + }, + { + "epoch": 67.80451127819549, + "grad_norm": 0.02062239684164524, + "learning_rate": 0.00024293235830743172, + "loss": 0.0005, + "step": 45090 + }, + { + "epoch": 67.81954887218045, + "grad_norm": 0.02473076619207859, + "learning_rate": 0.00024214731023046793, + "loss": 0.0005, + "step": 45100 + }, + { + "epoch": 67.83458646616542, + "grad_norm": 0.005499634891748428, + "learning_rate": 0.00024136233965638194, + "loss": 0.0004, + "step": 45110 + }, + { + "epoch": 67.84962406015038, + "grad_norm": 0.01819116808474064, + "learning_rate": 0.00024057745433251636, + "loss": 0.0004, + "step": 45120 + }, + { + "epoch": 67.86466165413533, + "grad_norm": 0.013280542567372322, + "learning_rate": 0.00023979266200537251, + "loss": 0.0004, + "step": 45130 + }, + { + "epoch": 67.8796992481203, + "grad_norm": 0.014933164231479168, + "learning_rate": 0.00023900797042053382, + "loss": 0.0006, + "step": 45140 + }, + { + "epoch": 67.89473684210526, + "grad_norm": 0.015526422299444675, + "learning_rate": 0.00023822338732258937, + "loss": 0.0004, + "step": 45150 + }, + { + "epoch": 67.90977443609023, + "grad_norm": 0.016882218420505524, + "learning_rate": 0.00023743892045505763, + "loss": 0.0004, + "step": 45160 + }, + { + "epoch": 67.92481203007519, + "grad_norm": 0.002431891392916441, + "learning_rate": 0.0002366545775603099, + "loss": 0.0004, + "step": 45170 + }, + { + "epoch": 67.93984962406014, + "grad_norm": 0.00942177139222622, + "learning_rate": 0.00023587036637949389, + "loss": 0.0003, + "step": 45180 + }, + { + "epoch": 67.95488721804512, + "grad_norm": 0.012570438906550407, + "learning_rate": 0.00023508629465245735, + "loss": 0.0006, + "step": 45190 + }, + { + "epoch": 67.96992481203007, + "grad_norm": 0.004532132297754288, + "learning_rate": 0.00023430237011767165, + "loss": 0.0003, + "step": 45200 + }, + { + "epoch": 67.98496240601504, + "grad_norm": 0.007650718558579683, + "learning_rate": 0.00023351860051215554, + "loss": 0.0004, + "step": 45210 + }, + { + "epoch": 68.0, + "grad_norm": 0.0010920488275587559, + "learning_rate": 0.00023273499357139853, + "loss": 0.0003, + "step": 45220 + }, + { + "epoch": 68.01503759398496, + "grad_norm": 0.018125634640455246, + "learning_rate": 0.00023195155702928483, + "loss": 0.0004, + "step": 45230 + }, + { + "epoch": 68.03007518796993, + "grad_norm": 0.021005718037486076, + "learning_rate": 0.00023116829861801686, + "loss": 0.0005, + "step": 45240 + }, + { + "epoch": 68.04511278195488, + "grad_norm": 0.01504642516374588, + "learning_rate": 0.0002303852260680388, + "loss": 0.0003, + "step": 45250 + }, + { + "epoch": 68.06015037593986, + "grad_norm": 0.010165783576667309, + "learning_rate": 0.00022960234710796062, + "loss": 0.0003, + "step": 45260 + }, + { + "epoch": 68.07518796992481, + "grad_norm": 0.004347332287579775, + "learning_rate": 0.00022881966946448166, + "loss": 0.0004, + "step": 45270 + }, + { + "epoch": 68.09022556390977, + "grad_norm": 0.021693730726838112, + "learning_rate": 0.00022803720086231422, + "loss": 0.0004, + "step": 45280 + }, + { + "epoch": 68.10526315789474, + "grad_norm": 0.013630317524075508, + "learning_rate": 0.0002272549490241075, + "loss": 0.0004, + "step": 45290 + }, + { + "epoch": 68.1203007518797, + "grad_norm": 0.021666934713721275, + "learning_rate": 0.00022647292167037142, + "loss": 0.0004, + "step": 45300 + }, + { + "epoch": 68.13533834586467, + "grad_norm": 0.017717469483613968, + "learning_rate": 0.00022569112651940016, + "loss": 0.0004, + "step": 45310 + }, + { + "epoch": 68.15037593984962, + "grad_norm": 0.005135768558830023, + "learning_rate": 0.00022490957128719626, + "loss": 0.0003, + "step": 45320 + }, + { + "epoch": 68.16541353383458, + "grad_norm": 0.007390771061182022, + "learning_rate": 0.00022412826368739438, + "loss": 0.0003, + "step": 45330 + }, + { + "epoch": 68.18045112781955, + "grad_norm": 0.014173122122883797, + "learning_rate": 0.00022334721143118502, + "loss": 0.0005, + "step": 45340 + }, + { + "epoch": 68.19548872180451, + "grad_norm": 0.014097933657467365, + "learning_rate": 0.00022256642222723868, + "loss": 0.0006, + "step": 45350 + }, + { + "epoch": 68.21052631578948, + "grad_norm": 0.004581925459206104, + "learning_rate": 0.00022178590378162956, + "loss": 0.0003, + "step": 45360 + }, + { + "epoch": 68.22556390977444, + "grad_norm": 0.013514547608792782, + "learning_rate": 0.00022100566379775965, + "loss": 0.0004, + "step": 45370 + }, + { + "epoch": 68.2406015037594, + "grad_norm": 0.006437893956899643, + "learning_rate": 0.00022022570997628254, + "loss": 0.0004, + "step": 45380 + }, + { + "epoch": 68.25563909774436, + "grad_norm": 0.003300102660432458, + "learning_rate": 0.00021944605001502761, + "loss": 0.0003, + "step": 45390 + }, + { + "epoch": 68.27067669172932, + "grad_norm": 0.011499815620481968, + "learning_rate": 0.00021866669160892392, + "loss": 0.0002, + "step": 45400 + }, + { + "epoch": 68.28571428571429, + "grad_norm": 0.009131526574492455, + "learning_rate": 0.00021788764244992426, + "loss": 0.0002, + "step": 45410 + }, + { + "epoch": 68.30075187969925, + "grad_norm": 0.018396493047475815, + "learning_rate": 0.00021710891022692937, + "loss": 0.0005, + "step": 45420 + }, + { + "epoch": 68.3157894736842, + "grad_norm": 0.012637213803827763, + "learning_rate": 0.00021633050262571187, + "loss": 0.0003, + "step": 45430 + }, + { + "epoch": 68.33082706766918, + "grad_norm": 0.013116303831338882, + "learning_rate": 0.0002155524273288405, + "loss": 0.0004, + "step": 45440 + }, + { + "epoch": 68.34586466165413, + "grad_norm": 0.013538711704313755, + "learning_rate": 0.00021477469201560434, + "loss": 0.0005, + "step": 45450 + }, + { + "epoch": 68.3609022556391, + "grad_norm": 0.020673442631959915, + "learning_rate": 0.00021399730436193694, + "loss": 0.0004, + "step": 45460 + }, + { + "epoch": 68.37593984962406, + "grad_norm": 0.016880374401807785, + "learning_rate": 0.00021322027204034063, + "loss": 0.0003, + "step": 45470 + }, + { + "epoch": 68.39097744360902, + "grad_norm": 0.023060057312250137, + "learning_rate": 0.00021244360271981073, + "loss": 0.0004, + "step": 45480 + }, + { + "epoch": 68.40601503759399, + "grad_norm": 0.011334877461194992, + "learning_rate": 0.0002116673040657598, + "loss": 0.0003, + "step": 45490 + }, + { + "epoch": 68.42105263157895, + "grad_norm": 0.00786332506686449, + "learning_rate": 0.00021089138373994224, + "loss": 0.0003, + "step": 45500 + }, + { + "epoch": 68.43609022556392, + "grad_norm": 0.013569455593824387, + "learning_rate": 0.00021011584940037838, + "loss": 0.0003, + "step": 45510 + }, + { + "epoch": 68.45112781954887, + "grad_norm": 0.01502404548227787, + "learning_rate": 0.0002093407087012791, + "loss": 0.0006, + "step": 45520 + }, + { + "epoch": 68.46616541353383, + "grad_norm": 0.00270978850312531, + "learning_rate": 0.00020856596929297007, + "loss": 0.0003, + "step": 45530 + }, + { + "epoch": 68.4812030075188, + "grad_norm": 0.035701680928468704, + "learning_rate": 0.00020779163882181655, + "loss": 0.0003, + "step": 45540 + }, + { + "epoch": 68.49624060150376, + "grad_norm": 0.023973694071173668, + "learning_rate": 0.00020701772493014758, + "loss": 0.0004, + "step": 45550 + }, + { + "epoch": 68.51127819548873, + "grad_norm": 0.013862264342606068, + "learning_rate": 0.00020624423525618098, + "loss": 0.0004, + "step": 45560 + }, + { + "epoch": 68.52631578947368, + "grad_norm": 0.009282819926738739, + "learning_rate": 0.00020547117743394743, + "loss": 0.0003, + "step": 45570 + }, + { + "epoch": 68.54135338345864, + "grad_norm": 0.006946340668946505, + "learning_rate": 0.00020469855909321564, + "loss": 0.0003, + "step": 45580 + }, + { + "epoch": 68.55639097744361, + "grad_norm": 0.005746881011873484, + "learning_rate": 0.00020392638785941665, + "loss": 0.0004, + "step": 45590 + }, + { + "epoch": 68.57142857142857, + "grad_norm": 0.0379464253783226, + "learning_rate": 0.0002031546713535688, + "loss": 0.0003, + "step": 45600 + }, + { + "epoch": 68.58646616541354, + "grad_norm": 0.022333946079015732, + "learning_rate": 0.00020238341719220254, + "loss": 0.0003, + "step": 45610 + }, + { + "epoch": 68.6015037593985, + "grad_norm": 0.016810264438390732, + "learning_rate": 0.00020161263298728495, + "loss": 0.0003, + "step": 45620 + }, + { + "epoch": 68.61654135338345, + "grad_norm": 0.004090671893209219, + "learning_rate": 0.00020084232634614503, + "loss": 0.0004, + "step": 45630 + }, + { + "epoch": 68.63157894736842, + "grad_norm": 0.015898995101451874, + "learning_rate": 0.00020007250487139827, + "loss": 0.0005, + "step": 45640 + }, + { + "epoch": 68.64661654135338, + "grad_norm": 0.015561041422188282, + "learning_rate": 0.00019930317616087195, + "loss": 0.0003, + "step": 45650 + }, + { + "epoch": 68.66165413533835, + "grad_norm": 0.015890102833509445, + "learning_rate": 0.00019853434780752973, + "loss": 0.0004, + "step": 45660 + }, + { + "epoch": 68.67669172932331, + "grad_norm": 0.011716490611433983, + "learning_rate": 0.00019776602739939714, + "loss": 0.0006, + "step": 45670 + }, + { + "epoch": 68.69172932330827, + "grad_norm": 0.011503184214234352, + "learning_rate": 0.0001969982225194864, + "loss": 0.0003, + "step": 45680 + }, + { + "epoch": 68.70676691729324, + "grad_norm": 0.006440093740820885, + "learning_rate": 0.00019623094074572173, + "loss": 0.0004, + "step": 45690 + }, + { + "epoch": 68.7218045112782, + "grad_norm": 0.004985397215932608, + "learning_rate": 0.00019546418965086444, + "loss": 0.0004, + "step": 45700 + }, + { + "epoch": 68.73684210526316, + "grad_norm": 0.02013504132628441, + "learning_rate": 0.00019469797680243827, + "loss": 0.0003, + "step": 45710 + }, + { + "epoch": 68.75187969924812, + "grad_norm": 0.01481330581009388, + "learning_rate": 0.00019393230976265475, + "loss": 0.0004, + "step": 45720 + }, + { + "epoch": 68.76691729323308, + "grad_norm": 0.007186059840023518, + "learning_rate": 0.00019316719608833844, + "loss": 0.0004, + "step": 45730 + }, + { + "epoch": 68.78195488721805, + "grad_norm": 0.012659070082008839, + "learning_rate": 0.00019240264333085245, + "loss": 0.0005, + "step": 45740 + }, + { + "epoch": 68.796992481203, + "grad_norm": 0.009326820261776447, + "learning_rate": 0.00019163865903602372, + "loss": 0.0005, + "step": 45750 + }, + { + "epoch": 68.81203007518798, + "grad_norm": 0.006875918712466955, + "learning_rate": 0.0001908752507440689, + "loss": 0.0003, + "step": 45760 + }, + { + "epoch": 68.82706766917293, + "grad_norm": 0.010382444597780704, + "learning_rate": 0.0001901124259895196, + "loss": 0.0004, + "step": 45770 + }, + { + "epoch": 68.84210526315789, + "grad_norm": 0.01603040285408497, + "learning_rate": 0.0001893501923011482, + "loss": 0.0004, + "step": 45780 + }, + { + "epoch": 68.85714285714286, + "grad_norm": 0.004476386588066816, + "learning_rate": 0.00018858855720189346, + "loss": 0.0003, + "step": 45790 + }, + { + "epoch": 68.87218045112782, + "grad_norm": 0.009257243946194649, + "learning_rate": 0.00018782752820878634, + "loss": 0.0003, + "step": 45800 + }, + { + "epoch": 68.88721804511279, + "grad_norm": 0.008520293049514294, + "learning_rate": 0.00018706711283287576, + "loss": 0.0003, + "step": 45810 + }, + { + "epoch": 68.90225563909775, + "grad_norm": 0.01385542657226324, + "learning_rate": 0.00018630731857915452, + "loss": 0.0004, + "step": 45820 + }, + { + "epoch": 68.9172932330827, + "grad_norm": 0.020168764516711235, + "learning_rate": 0.00018554815294648505, + "loss": 0.0004, + "step": 45830 + }, + { + "epoch": 68.93233082706767, + "grad_norm": 0.028927795588970184, + "learning_rate": 0.00018478962342752584, + "loss": 0.0003, + "step": 45840 + }, + { + "epoch": 68.94736842105263, + "grad_norm": 0.008192545734345913, + "learning_rate": 0.00018403173750865685, + "loss": 0.0003, + "step": 45850 + }, + { + "epoch": 68.9624060150376, + "grad_norm": 0.013087206520140171, + "learning_rate": 0.00018327450266990617, + "loss": 0.0002, + "step": 45860 + }, + { + "epoch": 68.97744360902256, + "grad_norm": 0.013421142473816872, + "learning_rate": 0.00018251792638487597, + "loss": 0.0005, + "step": 45870 + }, + { + "epoch": 68.99248120300751, + "grad_norm": 0.013940267264842987, + "learning_rate": 0.00018176201612066874, + "loss": 0.0002, + "step": 45880 + }, + { + "epoch": 69.00751879699249, + "grad_norm": 0.009932495653629303, + "learning_rate": 0.00018100677933781362, + "loss": 0.0004, + "step": 45890 + }, + { + "epoch": 69.02255639097744, + "grad_norm": 0.01071279775351286, + "learning_rate": 0.0001802522234901927, + "loss": 0.0003, + "step": 45900 + }, + { + "epoch": 69.0375939849624, + "grad_norm": 0.017572596669197083, + "learning_rate": 0.00017949835602496767, + "loss": 0.0004, + "step": 45910 + }, + { + "epoch": 69.05263157894737, + "grad_norm": 0.0012968970695510507, + "learning_rate": 0.00017874518438250596, + "loss": 0.0003, + "step": 45920 + }, + { + "epoch": 69.06766917293233, + "grad_norm": 0.006352079100906849, + "learning_rate": 0.0001779927159963078, + "loss": 0.0002, + "step": 45930 + }, + { + "epoch": 69.0827067669173, + "grad_norm": 0.011850732378661633, + "learning_rate": 0.0001772409582929321, + "loss": 0.0004, + "step": 45940 + }, + { + "epoch": 69.09774436090225, + "grad_norm": 0.006425502710044384, + "learning_rate": 0.00017648991869192405, + "loss": 0.0003, + "step": 45950 + }, + { + "epoch": 69.11278195488721, + "grad_norm": 0.007768154144287109, + "learning_rate": 0.00017573960460574132, + "loss": 0.0002, + "step": 45960 + }, + { + "epoch": 69.12781954887218, + "grad_norm": 0.008873769082129002, + "learning_rate": 0.00017499002343968097, + "loss": 0.0002, + "step": 45970 + }, + { + "epoch": 69.14285714285714, + "grad_norm": 0.008394557982683182, + "learning_rate": 0.00017424118259180656, + "loss": 0.0002, + "step": 45980 + }, + { + "epoch": 69.15789473684211, + "grad_norm": 0.011725624091923237, + "learning_rate": 0.00017349308945287484, + "loss": 0.0002, + "step": 45990 + }, + { + "epoch": 69.17293233082707, + "grad_norm": 0.008211090229451656, + "learning_rate": 0.00017274575140626317, + "loss": 0.0003, + "step": 46000 + }, + { + "epoch": 69.17293233082707, + "eval_cer": 0.01767248472945532, + "eval_loss": 0.11628076434135437, + "eval_runtime": 160.4652, + "eval_samples_per_second": 99.878, + "eval_steps_per_second": 0.785, + "eval_wer": 0.0657736473153038, + "step": 46000 + }, + { + "epoch": 69.18796992481202, + "grad_norm": 0.003420202061533928, + "learning_rate": 0.00017199917582789631, + "loss": 0.0002, + "step": 46010 + }, + { + "epoch": 69.203007518797, + "grad_norm": 0.010590638034045696, + "learning_rate": 0.00017125337008617387, + "loss": 0.0003, + "step": 46020 + }, + { + "epoch": 69.21804511278195, + "grad_norm": 0.004217818845063448, + "learning_rate": 0.00017050834154189732, + "loss": 0.0003, + "step": 46030 + }, + { + "epoch": 69.23308270676692, + "grad_norm": 0.003138753119856119, + "learning_rate": 0.00016976409754819767, + "loss": 0.0009, + "step": 46040 + }, + { + "epoch": 69.24812030075188, + "grad_norm": 0.014853217639029026, + "learning_rate": 0.0001690206454504627, + "loss": 0.0005, + "step": 46050 + }, + { + "epoch": 69.26315789473684, + "grad_norm": 0.010086318477988243, + "learning_rate": 0.00016827799258626442, + "loss": 0.0003, + "step": 46060 + }, + { + "epoch": 69.2781954887218, + "grad_norm": 0.009132744744420052, + "learning_rate": 0.00016753614628528678, + "loss": 0.0004, + "step": 46070 + }, + { + "epoch": 69.29323308270676, + "grad_norm": 0.017043599858880043, + "learning_rate": 0.00016679511386925337, + "loss": 0.0003, + "step": 46080 + }, + { + "epoch": 69.30827067669173, + "grad_norm": 0.007556526456028223, + "learning_rate": 0.00016605490265185485, + "loss": 0.0005, + "step": 46090 + }, + { + "epoch": 69.32330827067669, + "grad_norm": 0.003996487241238356, + "learning_rate": 0.00016531551993867715, + "loss": 0.0004, + "step": 46100 + }, + { + "epoch": 69.33834586466165, + "grad_norm": 0.020642543211579323, + "learning_rate": 0.00016457697302712918, + "loss": 0.0003, + "step": 46110 + }, + { + "epoch": 69.35338345864662, + "grad_norm": 0.013272679410874844, + "learning_rate": 0.00016383926920637078, + "loss": 0.0003, + "step": 46120 + }, + { + "epoch": 69.36842105263158, + "grad_norm": 0.017812160775065422, + "learning_rate": 0.00016310241575724077, + "loss": 0.0003, + "step": 46130 + }, + { + "epoch": 69.38345864661655, + "grad_norm": 0.010882627218961716, + "learning_rate": 0.0001623664199521853, + "loss": 0.0003, + "step": 46140 + }, + { + "epoch": 69.3984962406015, + "grad_norm": 0.01845810003578663, + "learning_rate": 0.00016163128905518576, + "loss": 0.0003, + "step": 46150 + }, + { + "epoch": 69.41353383458646, + "grad_norm": 0.013554912060499191, + "learning_rate": 0.00016089703032168734, + "loss": 0.0002, + "step": 46160 + }, + { + "epoch": 69.42857142857143, + "grad_norm": 0.030211202800273895, + "learning_rate": 0.00016016365099852736, + "loss": 0.0003, + "step": 46170 + }, + { + "epoch": 69.44360902255639, + "grad_norm": 0.014854599721729755, + "learning_rate": 0.0001594311583238636, + "loss": 0.0004, + "step": 46180 + }, + { + "epoch": 69.45864661654136, + "grad_norm": 0.006990060210227966, + "learning_rate": 0.00015869955952710308, + "loss": 0.0003, + "step": 46190 + }, + { + "epoch": 69.47368421052632, + "grad_norm": 0.008758706040680408, + "learning_rate": 0.00015796886182883053, + "loss": 0.0003, + "step": 46200 + }, + { + "epoch": 69.48872180451127, + "grad_norm": 0.016055459156632423, + "learning_rate": 0.0001572390724407373, + "loss": 0.0003, + "step": 46210 + }, + { + "epoch": 69.50375939849624, + "grad_norm": 0.01945725828409195, + "learning_rate": 0.00015651019856554994, + "loss": 0.0003, + "step": 46220 + }, + { + "epoch": 69.5187969924812, + "grad_norm": 0.026952844113111496, + "learning_rate": 0.00015578224739695937, + "loss": 0.0004, + "step": 46230 + }, + { + "epoch": 69.53383458646617, + "grad_norm": 0.00757815595716238, + "learning_rate": 0.00015505522611954976, + "loss": 0.0003, + "step": 46240 + }, + { + "epoch": 69.54887218045113, + "grad_norm": 0.015524659305810928, + "learning_rate": 0.00015432914190872756, + "loss": 0.0002, + "step": 46250 + }, + { + "epoch": 69.56390977443608, + "grad_norm": 0.00699151074513793, + "learning_rate": 0.00015360400193065087, + "loss": 0.0002, + "step": 46260 + }, + { + "epoch": 69.57894736842105, + "grad_norm": 0.010835791006684303, + "learning_rate": 0.00015287981334215851, + "loss": 0.0003, + "step": 46270 + }, + { + "epoch": 69.59398496240601, + "grad_norm": 0.016598382964730263, + "learning_rate": 0.0001521565832906994, + "loss": 0.0005, + "step": 46280 + }, + { + "epoch": 69.60902255639098, + "grad_norm": 0.00578779261559248, + "learning_rate": 0.00015143431891426223, + "loss": 0.0003, + "step": 46290 + }, + { + "epoch": 69.62406015037594, + "grad_norm": 0.009064720943570137, + "learning_rate": 0.00015071302734130488, + "loss": 0.0004, + "step": 46300 + }, + { + "epoch": 69.6390977443609, + "grad_norm": 0.015898587182164192, + "learning_rate": 0.00014999271569068385, + "loss": 0.0004, + "step": 46310 + }, + { + "epoch": 69.65413533834587, + "grad_norm": 0.011550568975508213, + "learning_rate": 0.00014927339107158436, + "loss": 0.0002, + "step": 46320 + }, + { + "epoch": 69.66917293233082, + "grad_norm": 0.012671023607254028, + "learning_rate": 0.00014855506058345002, + "loss": 0.0003, + "step": 46330 + }, + { + "epoch": 69.6842105263158, + "grad_norm": 0.006924602668732405, + "learning_rate": 0.00014783773131591278, + "loss": 0.0003, + "step": 46340 + }, + { + "epoch": 69.69924812030075, + "grad_norm": 0.02421189472079277, + "learning_rate": 0.0001471214103487228, + "loss": 0.0003, + "step": 46350 + }, + { + "epoch": 69.71428571428571, + "grad_norm": 0.017068902030587196, + "learning_rate": 0.00014640610475167898, + "loss": 0.0003, + "step": 46360 + }, + { + "epoch": 69.72932330827068, + "grad_norm": 0.004173495341092348, + "learning_rate": 0.00014569182158455873, + "loss": 0.0003, + "step": 46370 + }, + { + "epoch": 69.74436090225564, + "grad_norm": 0.006394834723323584, + "learning_rate": 0.00014497856789704843, + "loss": 0.0003, + "step": 46380 + }, + { + "epoch": 69.7593984962406, + "grad_norm": 0.011305994354188442, + "learning_rate": 0.00014426635072867423, + "loss": 0.0003, + "step": 46390 + }, + { + "epoch": 69.77443609022556, + "grad_norm": 0.022640669718384743, + "learning_rate": 0.00014355517710873183, + "loss": 0.0004, + "step": 46400 + }, + { + "epoch": 69.78947368421052, + "grad_norm": 0.02344363182783127, + "learning_rate": 0.00014284505405621795, + "loss": 0.0004, + "step": 46410 + }, + { + "epoch": 69.80451127819549, + "grad_norm": 0.0031091193668544292, + "learning_rate": 0.00014213598857976023, + "loss": 0.0003, + "step": 46420 + }, + { + "epoch": 69.81954887218045, + "grad_norm": 0.004734039772301912, + "learning_rate": 0.00014142798767754886, + "loss": 0.0003, + "step": 46430 + }, + { + "epoch": 69.83458646616542, + "grad_norm": 0.015176601707935333, + "learning_rate": 0.00014072105833726683, + "loss": 0.0003, + "step": 46440 + }, + { + "epoch": 69.84962406015038, + "grad_norm": 0.011872399598360062, + "learning_rate": 0.0001400152075360212, + "loss": 0.0002, + "step": 46450 + }, + { + "epoch": 69.86466165413533, + "grad_norm": 0.014066998846828938, + "learning_rate": 0.00013931044224027467, + "loss": 0.0003, + "step": 46460 + }, + { + "epoch": 69.8796992481203, + "grad_norm": 0.013536157086491585, + "learning_rate": 0.00013860676940577593, + "loss": 0.0003, + "step": 46470 + }, + { + "epoch": 69.89473684210526, + "grad_norm": 0.007990636862814426, + "learning_rate": 0.000137904195977492, + "loss": 0.0001, + "step": 46480 + }, + { + "epoch": 69.90977443609023, + "grad_norm": 0.009137459099292755, + "learning_rate": 0.0001372027288895387, + "loss": 0.0003, + "step": 46490 + }, + { + "epoch": 69.92481203007519, + "grad_norm": 0.009251121431589127, + "learning_rate": 0.00013650237506511331, + "loss": 0.0002, + "step": 46500 + }, + { + "epoch": 69.93984962406014, + "grad_norm": 0.0051016127690672874, + "learning_rate": 0.00013580314141642508, + "loss": 0.0003, + "step": 46510 + }, + { + "epoch": 69.95488721804512, + "grad_norm": 0.010840346105396748, + "learning_rate": 0.00013510503484462805, + "loss": 0.0003, + "step": 46520 + }, + { + "epoch": 69.96992481203007, + "grad_norm": 0.0025252399500459433, + "learning_rate": 0.0001344080622397521, + "loss": 0.0002, + "step": 46530 + }, + { + "epoch": 69.98496240601504, + "grad_norm": 0.008431227877736092, + "learning_rate": 0.00013371223048063541, + "loss": 0.0002, + "step": 46540 + }, + { + "epoch": 70.0, + "grad_norm": 0.015248334966599941, + "learning_rate": 0.0001330175464348567, + "loss": 0.0002, + "step": 46550 + }, + { + "epoch": 70.01503759398496, + "grad_norm": 0.0081748366355896, + "learning_rate": 0.00013232401695866685, + "loss": 0.0003, + "step": 46560 + }, + { + "epoch": 70.03007518796993, + "grad_norm": 0.016084836795926094, + "learning_rate": 0.00013163164889692198, + "loss": 0.0004, + "step": 46570 + }, + { + "epoch": 70.04511278195488, + "grad_norm": 0.004012197256088257, + "learning_rate": 0.0001309404490830152, + "loss": 0.0002, + "step": 46580 + }, + { + "epoch": 70.06015037593986, + "grad_norm": 0.01369634922593832, + "learning_rate": 0.00013025042433880977, + "loss": 0.0002, + "step": 46590 + }, + { + "epoch": 70.07518796992481, + "grad_norm": 0.015198171138763428, + "learning_rate": 0.00012956158147457115, + "loss": 0.0002, + "step": 46600 + }, + { + "epoch": 70.09022556390977, + "grad_norm": 0.008613619953393936, + "learning_rate": 0.00012887392728890053, + "loss": 0.0002, + "step": 46610 + }, + { + "epoch": 70.10526315789474, + "grad_norm": 0.015221030451357365, + "learning_rate": 0.00012818746856866687, + "loss": 0.0003, + "step": 46620 + }, + { + "epoch": 70.1203007518797, + "grad_norm": 0.014364690519869328, + "learning_rate": 0.00012750221208894085, + "loss": 0.0002, + "step": 46630 + }, + { + "epoch": 70.13533834586467, + "grad_norm": 0.013176646083593369, + "learning_rate": 0.00012681816461292713, + "loss": 0.0002, + "step": 46640 + }, + { + "epoch": 70.15037593984962, + "grad_norm": 0.012703269720077515, + "learning_rate": 0.0001261353328918981, + "loss": 0.0003, + "step": 46650 + }, + { + "epoch": 70.16541353383458, + "grad_norm": 0.014462068676948547, + "learning_rate": 0.0001254537236651273, + "loss": 0.0004, + "step": 46660 + }, + { + "epoch": 70.18045112781955, + "grad_norm": 0.015508248470723629, + "learning_rate": 0.00012477334365982248, + "loss": 0.0003, + "step": 46670 + }, + { + "epoch": 70.19548872180451, + "grad_norm": 0.011817226186394691, + "learning_rate": 0.0001240941995910598, + "loss": 0.0003, + "step": 46680 + }, + { + "epoch": 70.21052631578948, + "grad_norm": 0.004590926691889763, + "learning_rate": 0.0001234162981617168, + "loss": 0.0002, + "step": 46690 + }, + { + "epoch": 70.22556390977444, + "grad_norm": 0.0015648610424250364, + "learning_rate": 0.00012273964606240718, + "loss": 0.0003, + "step": 46700 + }, + { + "epoch": 70.2406015037594, + "grad_norm": 0.016540972515940666, + "learning_rate": 0.00012206424997141371, + "loss": 0.0004, + "step": 46710 + }, + { + "epoch": 70.25563909774436, + "grad_norm": 0.011143856681883335, + "learning_rate": 0.00012139011655462338, + "loss": 0.0002, + "step": 46720 + }, + { + "epoch": 70.27067669172932, + "grad_norm": 0.0110140610486269, + "learning_rate": 0.00012071725246546073, + "loss": 0.0003, + "step": 46730 + }, + { + "epoch": 70.28571428571429, + "grad_norm": 0.00558778690174222, + "learning_rate": 0.00012004566434482261, + "loss": 0.0002, + "step": 46740 + }, + { + "epoch": 70.30075187969925, + "grad_norm": 0.009393060579895973, + "learning_rate": 0.00011937535882101281, + "loss": 0.0003, + "step": 46750 + }, + { + "epoch": 70.3157894736842, + "grad_norm": 0.002385080559179187, + "learning_rate": 0.00011870634250967604, + "loss": 0.0003, + "step": 46760 + }, + { + "epoch": 70.33082706766918, + "grad_norm": 0.016264459118247032, + "learning_rate": 0.00011803862201373342, + "loss": 0.0002, + "step": 46770 + }, + { + "epoch": 70.34586466165413, + "grad_norm": 0.0189689751714468, + "learning_rate": 0.00011737220392331644, + "loss": 0.0002, + "step": 46780 + }, + { + "epoch": 70.3609022556391, + "grad_norm": 0.0021576143335551023, + "learning_rate": 0.00011670709481570285, + "loss": 0.0002, + "step": 46790 + }, + { + "epoch": 70.37593984962406, + "grad_norm": 0.02471758797764778, + "learning_rate": 0.00011604330125525078, + "loss": 0.0004, + "step": 46800 + }, + { + "epoch": 70.39097744360902, + "grad_norm": 0.0010209213942289352, + "learning_rate": 0.00011538082979333495, + "loss": 0.0002, + "step": 46810 + }, + { + "epoch": 70.40601503759399, + "grad_norm": 0.0022846111096441746, + "learning_rate": 0.00011471968696828106, + "loss": 0.0002, + "step": 46820 + }, + { + "epoch": 70.42105263157895, + "grad_norm": 0.021452682092785835, + "learning_rate": 0.00011405987930530184, + "loss": 0.0001, + "step": 46830 + }, + { + "epoch": 70.43609022556392, + "grad_norm": 0.018508970737457275, + "learning_rate": 0.00011340141331643275, + "loss": 0.0003, + "step": 46840 + }, + { + "epoch": 70.45112781954887, + "grad_norm": 0.013314683921635151, + "learning_rate": 0.00011274429550046702, + "loss": 0.0003, + "step": 46850 + }, + { + "epoch": 70.46616541353383, + "grad_norm": 0.014874989166855812, + "learning_rate": 0.00011208853234289245, + "loss": 0.0002, + "step": 46860 + }, + { + "epoch": 70.4812030075188, + "grad_norm": 0.0021699005737900734, + "learning_rate": 0.00011143413031582644, + "loss": 0.0002, + "step": 46870 + }, + { + "epoch": 70.49624060150376, + "grad_norm": 0.012988024391233921, + "learning_rate": 0.0001107810958779531, + "loss": 0.0003, + "step": 46880 + }, + { + "epoch": 70.51127819548873, + "grad_norm": 0.02212468720972538, + "learning_rate": 0.00011012943547445828, + "loss": 0.0004, + "step": 46890 + }, + { + "epoch": 70.52631578947368, + "grad_norm": 0.003696868196129799, + "learning_rate": 0.0001094791555369674, + "loss": 0.0001, + "step": 46900 + }, + { + "epoch": 70.54135338345864, + "grad_norm": 0.0054293442517519, + "learning_rate": 0.00010883026248348076, + "loss": 0.0003, + "step": 46910 + }, + { + "epoch": 70.55639097744361, + "grad_norm": 0.014588558115065098, + "learning_rate": 0.00010818276271831093, + "loss": 0.0003, + "step": 46920 + }, + { + "epoch": 70.57142857142857, + "grad_norm": 0.00038300349842756987, + "learning_rate": 0.00010753666263201906, + "loss": 0.0002, + "step": 46930 + }, + { + "epoch": 70.58646616541354, + "grad_norm": 0.004664332140237093, + "learning_rate": 0.00010689196860135234, + "loss": 0.0003, + "step": 46940 + }, + { + "epoch": 70.6015037593985, + "grad_norm": 0.004370580893009901, + "learning_rate": 0.00010624868698918044, + "loss": 0.0003, + "step": 46950 + }, + { + "epoch": 70.61654135338345, + "grad_norm": 0.003465394489467144, + "learning_rate": 0.00010560682414443315, + "loss": 0.0002, + "step": 46960 + }, + { + "epoch": 70.63157894736842, + "grad_norm": 0.006305689923465252, + "learning_rate": 0.00010496638640203774, + "loss": 0.0002, + "step": 46970 + }, + { + "epoch": 70.64661654135338, + "grad_norm": 0.006572013720870018, + "learning_rate": 0.00010432738008285602, + "loss": 0.0003, + "step": 46980 + }, + { + "epoch": 70.66165413533835, + "grad_norm": 0.0035556878428906202, + "learning_rate": 0.00010368981149362256, + "loss": 0.0003, + "step": 46990 + }, + { + "epoch": 70.67669172932331, + "grad_norm": 0.008983412757515907, + "learning_rate": 0.00010305368692688174, + "loss": 0.0002, + "step": 47000 + }, + { + "epoch": 70.67669172932331, + "eval_cer": 0.017787413344504305, + "eval_loss": 0.11624594777822495, + "eval_runtime": 159.2496, + "eval_samples_per_second": 100.641, + "eval_steps_per_second": 0.791, + "eval_wer": 0.0651628483903099, + "step": 47000 + }, + { + "epoch": 70.69172932330827, + "grad_norm": 0.007561844307929277, + "learning_rate": 0.00010241901266092644, + "loss": 0.0004, + "step": 47010 + }, + { + "epoch": 70.70676691729324, + "grad_norm": 0.014938493259251118, + "learning_rate": 0.0001017857949597352, + "loss": 0.0001, + "step": 47020 + }, + { + "epoch": 70.7218045112782, + "grad_norm": 0.003778475569561124, + "learning_rate": 0.00010115404007291131, + "loss": 0.0003, + "step": 47030 + }, + { + "epoch": 70.73684210526316, + "grad_norm": 0.0034889145754277706, + "learning_rate": 0.00010052375423562038, + "loss": 0.0002, + "step": 47040 + }, + { + "epoch": 70.75187969924812, + "grad_norm": 0.0014413069002330303, + "learning_rate": 9.989494366852902e-05, + "loss": 0.0002, + "step": 47050 + }, + { + "epoch": 70.76691729323308, + "grad_norm": 0.011321992613375187, + "learning_rate": 9.926761457774389e-05, + "loss": 0.0002, + "step": 47060 + }, + { + "epoch": 70.78195488721805, + "grad_norm": 0.0010370340896770358, + "learning_rate": 9.864177315474967e-05, + "loss": 0.0001, + "step": 47070 + }, + { + "epoch": 70.796992481203, + "grad_norm": 0.0057624028995633125, + "learning_rate": 9.801742557634872e-05, + "loss": 0.0003, + "step": 47080 + }, + { + "epoch": 70.81203007518798, + "grad_norm": 0.008389391005039215, + "learning_rate": 9.739457800459939e-05, + "loss": 0.0003, + "step": 47090 + }, + { + "epoch": 70.82706766917293, + "grad_norm": 0.004437154624611139, + "learning_rate": 9.677323658675594e-05, + "loss": 0.0002, + "step": 47100 + }, + { + "epoch": 70.84210526315789, + "grad_norm": 0.005877018440514803, + "learning_rate": 9.615340745520712e-05, + "loss": 0.0002, + "step": 47110 + }, + { + "epoch": 70.85714285714286, + "grad_norm": 0.0028633384499698877, + "learning_rate": 9.553509672741645e-05, + "loss": 0.0003, + "step": 47120 + }, + { + "epoch": 70.87218045112782, + "grad_norm": 0.01042623445391655, + "learning_rate": 9.491831050586108e-05, + "loss": 0.0003, + "step": 47130 + }, + { + "epoch": 70.88721804511279, + "grad_norm": 0.00323697691783309, + "learning_rate": 9.430305487797191e-05, + "loss": 0.0002, + "step": 47140 + }, + { + "epoch": 70.90225563909775, + "grad_norm": 0.00912206806242466, + "learning_rate": 9.368933591607378e-05, + "loss": 0.0002, + "step": 47150 + }, + { + "epoch": 70.9172932330827, + "grad_norm": 0.007355567999184132, + "learning_rate": 9.307715967732491e-05, + "loss": 0.0002, + "step": 47160 + }, + { + "epoch": 70.93233082706767, + "grad_norm": 0.006513164844363928, + "learning_rate": 9.246653220365778e-05, + "loss": 0.0002, + "step": 47170 + }, + { + "epoch": 70.94736842105263, + "grad_norm": 0.015151988714933395, + "learning_rate": 9.185745952171889e-05, + "loss": 0.0002, + "step": 47180 + }, + { + "epoch": 70.9624060150376, + "grad_norm": 0.012872877530753613, + "learning_rate": 9.124994764280989e-05, + "loss": 0.0002, + "step": 47190 + }, + { + "epoch": 70.97744360902256, + "grad_norm": 0.010012490674853325, + "learning_rate": 9.064400256282756e-05, + "loss": 0.0003, + "step": 47200 + }, + { + "epoch": 70.99248120300751, + "grad_norm": 0.004011506214737892, + "learning_rate": 9.003963026220543e-05, + "loss": 0.0002, + "step": 47210 + }, + { + "epoch": 71.00751879699249, + "grad_norm": 0.0030159244779497385, + "learning_rate": 8.94368367058539e-05, + "loss": 0.0001, + "step": 47220 + }, + { + "epoch": 71.02255639097744, + "grad_norm": 0.007837079465389252, + "learning_rate": 8.88356278431022e-05, + "loss": 0.0002, + "step": 47230 + }, + { + "epoch": 71.0375939849624, + "grad_norm": 0.005005802493542433, + "learning_rate": 8.8236009607639e-05, + "loss": 0.0002, + "step": 47240 + }, + { + "epoch": 71.05263157894737, + "grad_norm": 0.00966804102063179, + "learning_rate": 8.763798791745412e-05, + "loss": 0.0002, + "step": 47250 + }, + { + "epoch": 71.06766917293233, + "grad_norm": 0.0070493402890861034, + "learning_rate": 8.704156867478036e-05, + "loss": 0.0002, + "step": 47260 + }, + { + "epoch": 71.0827067669173, + "grad_norm": 0.015609405003488064, + "learning_rate": 8.644675776603475e-05, + "loss": 0.0002, + "step": 47270 + }, + { + "epoch": 71.09774436090225, + "grad_norm": 0.0012347523588687181, + "learning_rate": 8.585356106176093e-05, + "loss": 0.0001, + "step": 47280 + }, + { + "epoch": 71.11278195488721, + "grad_norm": 0.0017905740533024073, + "learning_rate": 8.526198441657077e-05, + "loss": 0.0001, + "step": 47290 + }, + { + "epoch": 71.12781954887218, + "grad_norm": 0.009683380834758282, + "learning_rate": 8.467203366908707e-05, + "loss": 0.0002, + "step": 47300 + }, + { + "epoch": 71.14285714285714, + "grad_norm": 0.011766936630010605, + "learning_rate": 8.408371464188536e-05, + "loss": 0.0003, + "step": 47310 + }, + { + "epoch": 71.15789473684211, + "grad_norm": 0.004445825237780809, + "learning_rate": 8.349703314143711e-05, + "loss": 0.0001, + "step": 47320 + }, + { + "epoch": 71.17293233082707, + "grad_norm": 0.0152448620647192, + "learning_rate": 8.29119949580518e-05, + "loss": 0.0002, + "step": 47330 + }, + { + "epoch": 71.18796992481202, + "grad_norm": 0.00861522275954485, + "learning_rate": 8.232860586582e-05, + "loss": 0.0001, + "step": 47340 + }, + { + "epoch": 71.203007518797, + "grad_norm": 0.005151827819645405, + "learning_rate": 8.174687162255672e-05, + "loss": 0.0002, + "step": 47350 + }, + { + "epoch": 71.21804511278195, + "grad_norm": 0.004266141448169947, + "learning_rate": 8.116679796974389e-05, + "loss": 0.0002, + "step": 47360 + }, + { + "epoch": 71.23308270676692, + "grad_norm": 0.0165784303098917, + "learning_rate": 8.058839063247447e-05, + "loss": 0.0003, + "step": 47370 + }, + { + "epoch": 71.24812030075188, + "grad_norm": 0.005976234097033739, + "learning_rate": 8.001165531939519e-05, + "loss": 0.0001, + "step": 47380 + }, + { + "epoch": 71.26315789473684, + "grad_norm": 0.011519176885485649, + "learning_rate": 7.94365977226509e-05, + "loss": 0.0002, + "step": 47390 + }, + { + "epoch": 71.2781954887218, + "grad_norm": 0.02166765369474888, + "learning_rate": 7.886322351782782e-05, + "loss": 0.0002, + "step": 47400 + }, + { + "epoch": 71.29323308270676, + "grad_norm": 0.008540966548025608, + "learning_rate": 7.829153836389796e-05, + "loss": 0.0001, + "step": 47410 + }, + { + "epoch": 71.30827067669173, + "grad_norm": 0.005044871475547552, + "learning_rate": 7.772154790316294e-05, + "loss": 0.0002, + "step": 47420 + }, + { + "epoch": 71.32330827067669, + "grad_norm": 0.006168388295918703, + "learning_rate": 7.715325776119841e-05, + "loss": 0.0002, + "step": 47430 + }, + { + "epoch": 71.33834586466165, + "grad_norm": 0.003097622888162732, + "learning_rate": 7.65866735467988e-05, + "loss": 0.0002, + "step": 47440 + }, + { + "epoch": 71.35338345864662, + "grad_norm": 0.019620204344391823, + "learning_rate": 7.602180085192142e-05, + "loss": 0.0003, + "step": 47450 + }, + { + "epoch": 71.36842105263158, + "grad_norm": 0.0037854290567338467, + "learning_rate": 7.545864525163188e-05, + "loss": 0.0002, + "step": 47460 + }, + { + "epoch": 71.38345864661655, + "grad_norm": 0.006471636239439249, + "learning_rate": 7.489721230404842e-05, + "loss": 0.0001, + "step": 47470 + }, + { + "epoch": 71.3984962406015, + "grad_norm": 0.006684000138193369, + "learning_rate": 7.433750755028773e-05, + "loss": 0.0002, + "step": 47480 + }, + { + "epoch": 71.41353383458646, + "grad_norm": 0.009500624611973763, + "learning_rate": 7.377953651440964e-05, + "loss": 0.0003, + "step": 47490 + }, + { + "epoch": 71.42857142857143, + "grad_norm": 0.0029275703709572554, + "learning_rate": 7.322330470336314e-05, + "loss": 0.0001, + "step": 47500 + }, + { + "epoch": 71.44360902255639, + "grad_norm": 0.007225681561976671, + "learning_rate": 7.266881760693158e-05, + "loss": 0.0002, + "step": 47510 + }, + { + "epoch": 71.45864661654136, + "grad_norm": 0.004247928969562054, + "learning_rate": 7.211608069767867e-05, + "loss": 0.0002, + "step": 47520 + }, + { + "epoch": 71.47368421052632, + "grad_norm": 0.014656057581305504, + "learning_rate": 7.156509943089471e-05, + "loss": 0.0002, + "step": 47530 + }, + { + "epoch": 71.48872180451127, + "grad_norm": 0.002657400444149971, + "learning_rate": 7.10158792445422e-05, + "loss": 0.0002, + "step": 47540 + }, + { + "epoch": 71.50375939849624, + "grad_norm": 0.0028791690710932016, + "learning_rate": 7.046842555920283e-05, + "loss": 0.0002, + "step": 47550 + }, + { + "epoch": 71.5187969924812, + "grad_norm": 0.003473178716376424, + "learning_rate": 6.992274377802327e-05, + "loss": 0.0002, + "step": 47560 + }, + { + "epoch": 71.53383458646617, + "grad_norm": 0.012211725115776062, + "learning_rate": 6.937883928666256e-05, + "loss": 0.0002, + "step": 47570 + }, + { + "epoch": 71.54887218045113, + "grad_norm": 0.0015056338161230087, + "learning_rate": 6.883671745323833e-05, + "loss": 0.0001, + "step": 47580 + }, + { + "epoch": 71.56390977443608, + "grad_norm": 0.00987530779093504, + "learning_rate": 6.829638362827431e-05, + "loss": 0.0003, + "step": 47590 + }, + { + "epoch": 71.57894736842105, + "grad_norm": 0.00797912199050188, + "learning_rate": 6.775784314464717e-05, + "loss": 0.0003, + "step": 47600 + }, + { + "epoch": 71.59398496240601, + "grad_norm": 0.0070827193558216095, + "learning_rate": 6.722110131753398e-05, + "loss": 0.0003, + "step": 47610 + }, + { + "epoch": 71.60902255639098, + "grad_norm": 0.0026458408683538437, + "learning_rate": 6.668616344436005e-05, + "loss": 0.0003, + "step": 47620 + }, + { + "epoch": 71.62406015037594, + "grad_norm": 0.0041486830450594425, + "learning_rate": 6.615303480474601e-05, + "loss": 0.0001, + "step": 47630 + }, + { + "epoch": 71.6390977443609, + "grad_norm": 0.005176181439310312, + "learning_rate": 6.562172066045655e-05, + "loss": 0.0002, + "step": 47640 + }, + { + "epoch": 71.65413533834587, + "grad_norm": 0.0015666828257963061, + "learning_rate": 6.509222625534755e-05, + "loss": 0.0001, + "step": 47650 + }, + { + "epoch": 71.66917293233082, + "grad_norm": 0.012206192128360271, + "learning_rate": 6.456455681531522e-05, + "loss": 0.0001, + "step": 47660 + }, + { + "epoch": 71.6842105263158, + "grad_norm": 0.019451946020126343, + "learning_rate": 6.403871754824372e-05, + "loss": 0.0003, + "step": 47670 + }, + { + "epoch": 71.69924812030075, + "grad_norm": 0.017617331817746162, + "learning_rate": 6.351471364395448e-05, + "loss": 0.0002, + "step": 47680 + }, + { + "epoch": 71.71428571428571, + "grad_norm": 0.0015220479108393192, + "learning_rate": 6.299255027415443e-05, + "loss": 0.0002, + "step": 47690 + }, + { + "epoch": 71.72932330827068, + "grad_norm": 0.009923718869686127, + "learning_rate": 6.24722325923851e-05, + "loss": 0.0002, + "step": 47700 + }, + { + "epoch": 71.74436090225564, + "grad_norm": 0.012326488271355629, + "learning_rate": 6.195376573397218e-05, + "loss": 0.0001, + "step": 47710 + }, + { + "epoch": 71.7593984962406, + "grad_norm": 0.0018185053486377, + "learning_rate": 6.143715481597403e-05, + "loss": 0.0001, + "step": 47720 + }, + { + "epoch": 71.77443609022556, + "grad_norm": 0.0016308833146467805, + "learning_rate": 6.0922404937132054e-05, + "loss": 0.0001, + "step": 47730 + }, + { + "epoch": 71.78947368421052, + "grad_norm": 0.01607769913971424, + "learning_rate": 6.040952117781953e-05, + "loss": 0.0002, + "step": 47740 + }, + { + "epoch": 71.80451127819549, + "grad_norm": 0.0022328388877213, + "learning_rate": 5.989850859999227e-05, + "loss": 0.0002, + "step": 47750 + }, + { + "epoch": 71.81954887218045, + "grad_norm": 0.009314276278018951, + "learning_rate": 5.9389372247138004e-05, + "loss": 0.0002, + "step": 47760 + }, + { + "epoch": 71.83458646616542, + "grad_norm": 0.003933258820325136, + "learning_rate": 5.8882117144227115e-05, + "loss": 0.0002, + "step": 47770 + }, + { + "epoch": 71.84962406015038, + "grad_norm": 0.00392696401104331, + "learning_rate": 5.837674829766257e-05, + "loss": 0.0001, + "step": 47780 + }, + { + "epoch": 71.86466165413533, + "grad_norm": 0.001757009536959231, + "learning_rate": 5.787327069523085e-05, + "loss": 0.0002, + "step": 47790 + }, + { + "epoch": 71.8796992481203, + "grad_norm": 0.0008779562776908278, + "learning_rate": 5.737168930605272e-05, + "loss": 0.0001, + "step": 47800 + }, + { + "epoch": 71.89473684210526, + "grad_norm": 0.007199555169790983, + "learning_rate": 5.6872009080533885e-05, + "loss": 0.0002, + "step": 47810 + }, + { + "epoch": 71.90977443609023, + "grad_norm": 0.011570352129638195, + "learning_rate": 5.637423495031657e-05, + "loss": 0.0001, + "step": 47820 + }, + { + "epoch": 71.92481203007519, + "grad_norm": 0.0008441720274277031, + "learning_rate": 5.587837182823033e-05, + "loss": 0.0001, + "step": 47830 + }, + { + "epoch": 71.93984962406014, + "grad_norm": 0.0023353747092187405, + "learning_rate": 5.5384424608244165e-05, + "loss": 0.0003, + "step": 47840 + }, + { + "epoch": 71.95488721804512, + "grad_norm": 0.012263569980859756, + "learning_rate": 5.489239816541755e-05, + "loss": 0.0001, + "step": 47850 + }, + { + "epoch": 71.96992481203007, + "grad_norm": 0.005669255740940571, + "learning_rate": 5.440229735585297e-05, + "loss": 0.0001, + "step": 47860 + }, + { + "epoch": 71.98496240601504, + "grad_norm": 0.009621870703995228, + "learning_rate": 5.391412701664744e-05, + "loss": 0.0002, + "step": 47870 + }, + { + "epoch": 72.0, + "grad_norm": 0.00277562253177166, + "learning_rate": 5.342789196584527e-05, + "loss": 0.0002, + "step": 47880 + }, + { + "epoch": 72.01503759398496, + "grad_norm": 0.019818391650915146, + "learning_rate": 5.294359700239001e-05, + "loss": 0.0002, + "step": 47890 + }, + { + "epoch": 72.03007518796993, + "grad_norm": 0.0019190360326319933, + "learning_rate": 5.2461246906077396e-05, + "loss": 0.0001, + "step": 47900 + }, + { + "epoch": 72.04511278195488, + "grad_norm": 0.005672231782227755, + "learning_rate": 5.198084643750825e-05, + "loss": 0.0002, + "step": 47910 + }, + { + "epoch": 72.06015037593986, + "grad_norm": 0.00741492910310626, + "learning_rate": 5.1502400338041156e-05, + "loss": 0.0002, + "step": 47920 + }, + { + "epoch": 72.07518796992481, + "grad_norm": 0.0010002661729231477, + "learning_rate": 5.102591332974604e-05, + "loss": 0.0001, + "step": 47930 + }, + { + "epoch": 72.09022556390977, + "grad_norm": 0.007011167239397764, + "learning_rate": 5.0551390115357225e-05, + "loss": 0.0001, + "step": 47940 + }, + { + "epoch": 72.10526315789474, + "grad_norm": 0.0030187827069312334, + "learning_rate": 5.007883537822736e-05, + "loss": 0.0002, + "step": 47950 + }, + { + "epoch": 72.1203007518797, + "grad_norm": 0.014452381059527397, + "learning_rate": 4.960825378228082e-05, + "loss": 0.0001, + "step": 47960 + }, + { + "epoch": 72.13533834586467, + "grad_norm": 0.006032969802618027, + "learning_rate": 4.91396499719681e-05, + "loss": 0.0002, + "step": 47970 + }, + { + "epoch": 72.15037593984962, + "grad_norm": 0.007046033628284931, + "learning_rate": 4.867302857221953e-05, + "loss": 0.0002, + "step": 47980 + }, + { + "epoch": 72.16541353383458, + "grad_norm": 0.007812702096998692, + "learning_rate": 4.820839418839992e-05, + "loss": 0.0001, + "step": 47990 + }, + { + "epoch": 72.18045112781955, + "grad_norm": 0.006637748796492815, + "learning_rate": 4.7745751406263163e-05, + "loss": 0.0001, + "step": 48000 + }, + { + "epoch": 72.18045112781955, + "eval_cer": 0.017652763251133304, + "eval_loss": 0.122411347925663, + "eval_runtime": 160.2594, + "eval_samples_per_second": 100.007, + "eval_steps_per_second": 0.786, + "eval_wer": 0.06506887932492623, + "step": 48000 + }, + { + "epoch": 72.19548872180451, + "grad_norm": 0.008753658272325993, + "learning_rate": 4.7285104791906617e-05, + "loss": 0.0002, + "step": 48010 + }, + { + "epoch": 72.21052631578948, + "grad_norm": 0.003546468447893858, + "learning_rate": 4.6826458891726513e-05, + "loss": 0.0001, + "step": 48020 + }, + { + "epoch": 72.22556390977444, + "grad_norm": 0.01917950250208378, + "learning_rate": 4.636981823237263e-05, + "loss": 0.0003, + "step": 48030 + }, + { + "epoch": 72.2406015037594, + "grad_norm": 0.00047616424853913486, + "learning_rate": 4.5915187320704016e-05, + "loss": 0.0002, + "step": 48040 + }, + { + "epoch": 72.25563909774436, + "grad_norm": 0.013119297102093697, + "learning_rate": 4.54625706437441e-05, + "loss": 0.0002, + "step": 48050 + }, + { + "epoch": 72.27067669172932, + "grad_norm": 0.012134304270148277, + "learning_rate": 4.501197266863691e-05, + "loss": 0.0001, + "step": 48060 + }, + { + "epoch": 72.28571428571429, + "grad_norm": 0.0006046507623977959, + "learning_rate": 4.456339784260246e-05, + "loss": 0.0002, + "step": 48070 + }, + { + "epoch": 72.30075187969925, + "grad_norm": 0.013165694661438465, + "learning_rate": 4.411685059289314e-05, + "loss": 0.0002, + "step": 48080 + }, + { + "epoch": 72.3157894736842, + "grad_norm": 0.009810341522097588, + "learning_rate": 4.367233532675011e-05, + "loss": 0.0001, + "step": 48090 + }, + { + "epoch": 72.33082706766918, + "grad_norm": 0.0009752390324138105, + "learning_rate": 4.3229856431359515e-05, + "loss": 0.0002, + "step": 48100 + }, + { + "epoch": 72.34586466165413, + "grad_norm": 0.008896918036043644, + "learning_rate": 4.278941827380953e-05, + "loss": 0.0002, + "step": 48110 + }, + { + "epoch": 72.3609022556391, + "grad_norm": 0.0021772964391857386, + "learning_rate": 4.2351025201046804e-05, + "loss": 0.0001, + "step": 48120 + }, + { + "epoch": 72.37593984962406, + "grad_norm": 0.0012971338583156466, + "learning_rate": 4.191468153983419e-05, + "loss": 0.0002, + "step": 48130 + }, + { + "epoch": 72.39097744360902, + "grad_norm": 0.008249860256910324, + "learning_rate": 4.148039159670722e-05, + "loss": 0.0001, + "step": 48140 + }, + { + "epoch": 72.40601503759399, + "grad_norm": 0.0007850232068449259, + "learning_rate": 4.104815965793249e-05, + "loss": 0.0001, + "step": 48150 + }, + { + "epoch": 72.42105263157895, + "grad_norm": 0.015763426199555397, + "learning_rate": 4.0617989989464586e-05, + "loss": 0.0002, + "step": 48160 + }, + { + "epoch": 72.43609022556392, + "grad_norm": 0.011626193299889565, + "learning_rate": 4.018988683690461e-05, + "loss": 0.0001, + "step": 48170 + }, + { + "epoch": 72.45112781954887, + "grad_norm": 0.011490133590996265, + "learning_rate": 3.976385442545774e-05, + "loss": 0.0001, + "step": 48180 + }, + { + "epoch": 72.46616541353383, + "grad_norm": 0.0017297123558819294, + "learning_rate": 3.9339896959891985e-05, + "loss": 0.0002, + "step": 48190 + }, + { + "epoch": 72.4812030075188, + "grad_norm": 0.0022541387006640434, + "learning_rate": 3.891801862449629e-05, + "loss": 0.0001, + "step": 48200 + }, + { + "epoch": 72.49624060150376, + "grad_norm": 0.009118613786995411, + "learning_rate": 3.8498223583039476e-05, + "loss": 0.0002, + "step": 48210 + }, + { + "epoch": 72.51127819548873, + "grad_norm": 0.004675202537328005, + "learning_rate": 3.808051597872925e-05, + "loss": 0.0001, + "step": 48220 + }, + { + "epoch": 72.52631578947368, + "grad_norm": 0.06415645778179169, + "learning_rate": 3.766489993417088e-05, + "loss": 0.0002, + "step": 48230 + }, + { + "epoch": 72.54135338345864, + "grad_norm": 0.019507089629769325, + "learning_rate": 3.725137955132707e-05, + "loss": 0.0001, + "step": 48240 + }, + { + "epoch": 72.55639097744361, + "grad_norm": 0.010918454267084599, + "learning_rate": 3.6839958911476953e-05, + "loss": 0.0001, + "step": 48250 + }, + { + "epoch": 72.57142857142857, + "grad_norm": 0.015231043100357056, + "learning_rate": 3.643064207517624e-05, + "loss": 0.0002, + "step": 48260 + }, + { + "epoch": 72.58646616541354, + "grad_norm": 0.01048593781888485, + "learning_rate": 3.602343308221675e-05, + "loss": 0.0002, + "step": 48270 + }, + { + "epoch": 72.6015037593985, + "grad_norm": 0.00857542734593153, + "learning_rate": 3.561833595158698e-05, + "loss": 0.0002, + "step": 48280 + }, + { + "epoch": 72.61654135338345, + "grad_norm": 0.003248027293011546, + "learning_rate": 3.521535468143197e-05, + "loss": 0.0002, + "step": 48290 + }, + { + "epoch": 72.63157894736842, + "grad_norm": 0.0027482425794005394, + "learning_rate": 3.481449324901412e-05, + "loss": 0.0001, + "step": 48300 + }, + { + "epoch": 72.64661654135338, + "grad_norm": 0.005018073134124279, + "learning_rate": 3.441575561067406e-05, + "loss": 0.0002, + "step": 48310 + }, + { + "epoch": 72.66165413533835, + "grad_norm": 0.0029400414787232876, + "learning_rate": 3.401914570179118e-05, + "loss": 0.0001, + "step": 48320 + }, + { + "epoch": 72.67669172932331, + "grad_norm": 0.018955707550048828, + "learning_rate": 3.3624667436745305e-05, + "loss": 0.0002, + "step": 48330 + }, + { + "epoch": 72.69172932330827, + "grad_norm": 0.005142426583915949, + "learning_rate": 3.323232470887749e-05, + "loss": 0.0002, + "step": 48340 + }, + { + "epoch": 72.70676691729324, + "grad_norm": 0.002180666895583272, + "learning_rate": 3.284212139045223e-05, + "loss": 0.0001, + "step": 48350 + }, + { + "epoch": 72.7218045112782, + "grad_norm": 0.0013281555147841573, + "learning_rate": 3.245406133261858e-05, + "loss": 0.0001, + "step": 48360 + }, + { + "epoch": 72.73684210526316, + "grad_norm": 0.0019101777579635382, + "learning_rate": 3.206814836537281e-05, + "loss": 0.0001, + "step": 48370 + }, + { + "epoch": 72.75187969924812, + "grad_norm": 0.008839000016450882, + "learning_rate": 3.168438629752002e-05, + "loss": 0.0001, + "step": 48380 + }, + { + "epoch": 72.76691729323308, + "grad_norm": 0.0008121269638650119, + "learning_rate": 3.1302778916636824e-05, + "loss": 0.0001, + "step": 48390 + }, + { + "epoch": 72.78195488721805, + "grad_norm": 0.03309241682291031, + "learning_rate": 3.092332998903416e-05, + "loss": 0.0002, + "step": 48400 + }, + { + "epoch": 72.796992481203, + "grad_norm": 0.010003205388784409, + "learning_rate": 3.05460432597196e-05, + "loss": 0.0002, + "step": 48410 + }, + { + "epoch": 72.81203007518798, + "grad_norm": 0.026142679154872894, + "learning_rate": 3.017092245236097e-05, + "loss": 0.0002, + "step": 48420 + }, + { + "epoch": 72.82706766917293, + "grad_norm": 0.007639321964234114, + "learning_rate": 2.9797971269249103e-05, + "loss": 0.0002, + "step": 48430 + }, + { + "epoch": 72.84210526315789, + "grad_norm": 0.004819975234568119, + "learning_rate": 2.942719339126171e-05, + "loss": 0.0001, + "step": 48440 + }, + { + "epoch": 72.85714285714286, + "grad_norm": 0.00519084045663476, + "learning_rate": 2.9058592477826635e-05, + "loss": 0.0002, + "step": 48450 + }, + { + "epoch": 72.87218045112782, + "grad_norm": 0.006006615236401558, + "learning_rate": 2.8692172166886215e-05, + "loss": 0.0001, + "step": 48460 + }, + { + "epoch": 72.88721804511279, + "grad_norm": 0.00965845212340355, + "learning_rate": 2.8327936074860865e-05, + "loss": 0.0003, + "step": 48470 + }, + { + "epoch": 72.90225563909775, + "grad_norm": 0.005012243054807186, + "learning_rate": 2.796588779661388e-05, + "loss": 0.0001, + "step": 48480 + }, + { + "epoch": 72.9172932330827, + "grad_norm": 0.003911204636096954, + "learning_rate": 2.7606030905415552e-05, + "loss": 0.0001, + "step": 48490 + }, + { + "epoch": 72.93233082706767, + "grad_norm": 0.013586202636361122, + "learning_rate": 2.7248368952908055e-05, + "loss": 0.0002, + "step": 48500 + }, + { + "epoch": 72.94736842105263, + "grad_norm": 0.000793063489254564, + "learning_rate": 2.6892905469070554e-05, + "loss": 0.0002, + "step": 48510 + }, + { + "epoch": 72.9624060150376, + "grad_norm": 0.006107704248279333, + "learning_rate": 2.6539643962184058e-05, + "loss": 0.0001, + "step": 48520 + }, + { + "epoch": 72.97744360902256, + "grad_norm": 0.011627903208136559, + "learning_rate": 2.618858791879711e-05, + "loss": 0.0002, + "step": 48530 + }, + { + "epoch": 72.99248120300751, + "grad_norm": 0.006152286194264889, + "learning_rate": 2.5839740803691032e-05, + "loss": 0.0002, + "step": 48540 + }, + { + "epoch": 73.00751879699249, + "grad_norm": 0.0003537592419888824, + "learning_rate": 2.5493106059846115e-05, + "loss": 0.0001, + "step": 48550 + }, + { + "epoch": 73.02255639097744, + "grad_norm": 0.0012257678899914026, + "learning_rate": 2.514868710840723e-05, + "loss": 0.0001, + "step": 48560 + }, + { + "epoch": 73.0375939849624, + "grad_norm": 0.007412443868815899, + "learning_rate": 2.4806487348650486e-05, + "loss": 0.0002, + "step": 48570 + }, + { + "epoch": 73.05263157894737, + "grad_norm": 0.006936053745448589, + "learning_rate": 2.4466510157949318e-05, + "loss": 0.0001, + "step": 48580 + }, + { + "epoch": 73.06766917293233, + "grad_norm": 0.009363191202282906, + "learning_rate": 2.412875889174129e-05, + "loss": 0.0001, + "step": 48590 + }, + { + "epoch": 73.0827067669173, + "grad_norm": 0.007069251034408808, + "learning_rate": 2.379323688349516e-05, + "loss": 0.0001, + "step": 48600 + }, + { + "epoch": 73.09774436090225, + "grad_norm": 0.006997474003583193, + "learning_rate": 2.3459947444677553e-05, + "loss": 0.0001, + "step": 48610 + }, + { + "epoch": 73.11278195488721, + "grad_norm": 0.0017135557718575, + "learning_rate": 2.312889386472078e-05, + "loss": 0.0002, + "step": 48620 + }, + { + "epoch": 73.12781954887218, + "grad_norm": 0.0005650724051520228, + "learning_rate": 2.2800079410989966e-05, + "loss": 0.0001, + "step": 48630 + }, + { + "epoch": 73.14285714285714, + "grad_norm": 0.004696905612945557, + "learning_rate": 2.2473507328751085e-05, + "loss": 0.0002, + "step": 48640 + }, + { + "epoch": 73.15789473684211, + "grad_norm": 0.008275547996163368, + "learning_rate": 2.214918084113868e-05, + "loss": 0.0001, + "step": 48650 + }, + { + "epoch": 73.17293233082707, + "grad_norm": 0.0038258659187704325, + "learning_rate": 2.1827103149124312e-05, + "loss": 0.0001, + "step": 48660 + }, + { + "epoch": 73.18796992481202, + "grad_norm": 0.0038346555083990097, + "learning_rate": 2.150727743148473e-05, + "loss": 0.0001, + "step": 48670 + }, + { + "epoch": 73.203007518797, + "grad_norm": 0.0006511756801046431, + "learning_rate": 2.1189706844770618e-05, + "loss": 0.0001, + "step": 48680 + }, + { + "epoch": 73.21804511278195, + "grad_norm": 0.004391991999000311, + "learning_rate": 2.0874394523275526e-05, + "loss": 0.0001, + "step": 48690 + }, + { + "epoch": 73.23308270676692, + "grad_norm": 0.008018497377634048, + "learning_rate": 2.0561343579004716e-05, + "loss": 0.0002, + "step": 48700 + }, + { + "epoch": 73.24812030075188, + "grad_norm": 0.006681958679109812, + "learning_rate": 2.0250557101644697e-05, + "loss": 0.0002, + "step": 48710 + }, + { + "epoch": 73.26315789473684, + "grad_norm": 0.008693943731486797, + "learning_rate": 1.9942038158532405e-05, + "loss": 0.0001, + "step": 48720 + }, + { + "epoch": 73.2781954887218, + "grad_norm": 0.014171193353831768, + "learning_rate": 1.963578979462541e-05, + "loss": 0.0003, + "step": 48730 + }, + { + "epoch": 73.29323308270676, + "grad_norm": 0.004971709568053484, + "learning_rate": 1.9331815032471277e-05, + "loss": 0.0002, + "step": 48740 + }, + { + "epoch": 73.30827067669173, + "grad_norm": 0.008193280547857285, + "learning_rate": 1.9030116872178316e-05, + "loss": 0.0001, + "step": 48750 + }, + { + "epoch": 73.32330827067669, + "grad_norm": 0.00471587385982275, + "learning_rate": 1.873069829138552e-05, + "loss": 0.0002, + "step": 48760 + }, + { + "epoch": 73.33834586466165, + "grad_norm": 0.020396694540977478, + "learning_rate": 1.843356224523335e-05, + "loss": 0.0003, + "step": 48770 + }, + { + "epoch": 73.35338345864662, + "grad_norm": 0.006142711732536554, + "learning_rate": 1.8138711666334683e-05, + "loss": 0.0002, + "step": 48780 + }, + { + "epoch": 73.36842105263158, + "grad_norm": 0.01006547175347805, + "learning_rate": 1.7846149464745666e-05, + "loss": 0.0002, + "step": 48790 + }, + { + "epoch": 73.38345864661655, + "grad_norm": 0.010062182322144508, + "learning_rate": 1.7555878527937163e-05, + "loss": 0.0002, + "step": 48800 + }, + { + "epoch": 73.3984962406015, + "grad_norm": 0.009443131275475025, + "learning_rate": 1.726790172076606e-05, + "loss": 0.0002, + "step": 48810 + }, + { + "epoch": 73.41353383458646, + "grad_norm": 0.008163406513631344, + "learning_rate": 1.6982221885447263e-05, + "loss": 0.0001, + "step": 48820 + }, + { + "epoch": 73.42857142857143, + "grad_norm": 0.008576576597988605, + "learning_rate": 1.669884184152534e-05, + "loss": 0.0001, + "step": 48830 + }, + { + "epoch": 73.44360902255639, + "grad_norm": 0.002591648604720831, + "learning_rate": 1.6417764385846996e-05, + "loss": 0.0001, + "step": 48840 + }, + { + "epoch": 73.45864661654136, + "grad_norm": 0.002301494823768735, + "learning_rate": 1.6138992292533183e-05, + "loss": 0.0001, + "step": 48850 + }, + { + "epoch": 73.47368421052632, + "grad_norm": 0.0123626459389925, + "learning_rate": 1.586252831295193e-05, + "loss": 0.0001, + "step": 48860 + }, + { + "epoch": 73.48872180451127, + "grad_norm": 0.001510324073024094, + "learning_rate": 1.5588375175691116e-05, + "loss": 0.0002, + "step": 48870 + }, + { + "epoch": 73.50375939849624, + "grad_norm": 0.009397964924573898, + "learning_rate": 1.5316535586531482e-05, + "loss": 0.0002, + "step": 48880 + }, + { + "epoch": 73.5187969924812, + "grad_norm": 0.007383540738373995, + "learning_rate": 1.5047012228420088e-05, + "loss": 0.0002, + "step": 48890 + }, + { + "epoch": 73.53383458646617, + "grad_norm": 0.021646033972501755, + "learning_rate": 1.4779807761443637e-05, + "loss": 0.0001, + "step": 48900 + }, + { + "epoch": 73.54887218045113, + "grad_norm": 0.005388366524130106, + "learning_rate": 1.4514924822802367e-05, + "loss": 0.0002, + "step": 48910 + }, + { + "epoch": 73.56390977443608, + "grad_norm": 0.001189828384667635, + "learning_rate": 1.425236602678387e-05, + "loss": 0.0002, + "step": 48920 + }, + { + "epoch": 73.57894736842105, + "grad_norm": 0.012560972012579441, + "learning_rate": 1.3992133964737585e-05, + "loss": 0.0002, + "step": 48930 + }, + { + "epoch": 73.59398496240601, + "grad_norm": 0.004339755512773991, + "learning_rate": 1.3734231205048826e-05, + "loss": 0.0001, + "step": 48940 + }, + { + "epoch": 73.60902255639098, + "grad_norm": 0.0025037273298949003, + "learning_rate": 1.3478660293113675e-05, + "loss": 0.0001, + "step": 48950 + }, + { + "epoch": 73.62406015037594, + "grad_norm": 0.0032953948248177767, + "learning_rate": 1.3225423751313942e-05, + "loss": 0.0001, + "step": 48960 + }, + { + "epoch": 73.6390977443609, + "grad_norm": 0.026934852823615074, + "learning_rate": 1.2974524078991995e-05, + "loss": 0.0001, + "step": 48970 + }, + { + "epoch": 73.65413533834587, + "grad_norm": 0.011640683747828007, + "learning_rate": 1.2725963752426379e-05, + "loss": 0.0001, + "step": 48980 + }, + { + "epoch": 73.66917293233082, + "grad_norm": 0.009896302595734596, + "learning_rate": 1.2479745224807049e-05, + "loss": 0.0001, + "step": 48990 + }, + { + "epoch": 73.6842105263158, + "grad_norm": 0.005009777843952179, + "learning_rate": 1.2235870926211617e-05, + "loss": 0.0002, + "step": 49000 + }, + { + "epoch": 73.6842105263158, + "eval_cer": 0.017637802129647637, + "eval_loss": 0.12418342381715775, + "eval_runtime": 161.7974, + "eval_samples_per_second": 99.056, + "eval_steps_per_second": 0.779, + "eval_wer": 0.06516754684357909, + "step": 49000 + }, + { + "epoch": 73.69924812030075, + "grad_norm": 0.010720503516495228, + "learning_rate": 1.1994343263580843e-05, + "loss": 0.0002, + "step": 49010 + }, + { + "epoch": 73.71428571428571, + "grad_norm": 0.011719441041350365, + "learning_rate": 1.1755164620695314e-05, + "loss": 0.0002, + "step": 49020 + }, + { + "epoch": 73.72932330827068, + "grad_norm": 0.0016286182217299938, + "learning_rate": 1.1518337358151636e-05, + "loss": 0.0001, + "step": 49030 + }, + { + "epoch": 73.74436090225564, + "grad_norm": 0.00315404892899096, + "learning_rate": 1.1283863813339262e-05, + "loss": 0.0003, + "step": 49040 + }, + { + "epoch": 73.7593984962406, + "grad_norm": 0.005642259493470192, + "learning_rate": 1.105174630041747e-05, + "loss": 0.0002, + "step": 49050 + }, + { + "epoch": 73.77443609022556, + "grad_norm": 0.01121502835303545, + "learning_rate": 1.0821987110292364e-05, + "loss": 0.0001, + "step": 49060 + }, + { + "epoch": 73.78947368421052, + "grad_norm": 0.007396548055112362, + "learning_rate": 1.0594588510594445e-05, + "loss": 0.0001, + "step": 49070 + }, + { + "epoch": 73.80451127819549, + "grad_norm": 0.0017339944606646895, + "learning_rate": 1.0369552745656014e-05, + "loss": 0.0001, + "step": 49080 + }, + { + "epoch": 73.81954887218045, + "grad_norm": 0.004210019484162331, + "learning_rate": 1.0146882036489307e-05, + "loss": 0.0001, + "step": 49090 + }, + { + "epoch": 73.83458646616542, + "grad_norm": 0.001912950538098812, + "learning_rate": 9.926578580764234e-06, + "loss": 0.0002, + "step": 49100 + }, + { + "epoch": 73.84962406015038, + "grad_norm": 0.0026681837625801563, + "learning_rate": 9.708644552787028e-06, + "loss": 0.0002, + "step": 49110 + }, + { + "epoch": 73.86466165413533, + "grad_norm": 0.001706672483123839, + "learning_rate": 9.493082103478518e-06, + "loss": 0.0001, + "step": 49120 + }, + { + "epoch": 73.8796992481203, + "grad_norm": 0.0009129999671131372, + "learning_rate": 9.279893360353093e-06, + "loss": 0.0001, + "step": 49130 + }, + { + "epoch": 73.89473684210526, + "grad_norm": 0.0006233818712644279, + "learning_rate": 9.069080427497572e-06, + "loss": 0.0002, + "step": 49140 + }, + { + "epoch": 73.90977443609023, + "grad_norm": 0.0022554509341716766, + "learning_rate": 8.860645385550481e-06, + "loss": 0.0003, + "step": 49150 + }, + { + "epoch": 73.92481203007519, + "grad_norm": 0.005981481168419123, + "learning_rate": 8.65459029168153e-06, + "loss": 0.0002, + "step": 49160 + }, + { + "epoch": 73.93984962406014, + "grad_norm": 0.005019905511289835, + "learning_rate": 8.450917179571306e-06, + "loss": 0.0001, + "step": 49170 + }, + { + "epoch": 73.95488721804512, + "grad_norm": 0.004454623442143202, + "learning_rate": 8.249628059391251e-06, + "loss": 0.0001, + "step": 49180 + }, + { + "epoch": 73.96992481203007, + "grad_norm": 0.008569353260099888, + "learning_rate": 8.050724917783635e-06, + "loss": 0.0002, + "step": 49190 + }, + { + "epoch": 73.98496240601504, + "grad_norm": 0.015535306185483932, + "learning_rate": 7.854209717842232e-06, + "loss": 0.0002, + "step": 49200 + }, + { + "epoch": 74.0, + "grad_norm": 0.0011461131507530808, + "learning_rate": 7.660084399092659e-06, + "loss": 0.0002, + "step": 49210 + }, + { + "epoch": 74.01503759398496, + "grad_norm": 0.02057073451578617, + "learning_rate": 7.468350877473551e-06, + "loss": 0.0002, + "step": 49220 + }, + { + "epoch": 74.03007518796993, + "grad_norm": 0.004122065845876932, + "learning_rate": 7.279011045317252e-06, + "loss": 0.0001, + "step": 49230 + }, + { + "epoch": 74.04511278195488, + "grad_norm": 0.0026912291068583727, + "learning_rate": 7.092066771331507e-06, + "loss": 0.0002, + "step": 49240 + }, + { + "epoch": 74.06015037593986, + "grad_norm": 0.0011043765116482973, + "learning_rate": 6.907519900580861e-06, + "loss": 0.0002, + "step": 49250 + }, + { + "epoch": 74.07518796992481, + "grad_norm": 0.0006450146320275962, + "learning_rate": 6.725372254468343e-06, + "loss": 0.0001, + "step": 49260 + }, + { + "epoch": 74.09022556390977, + "grad_norm": 0.0015586280496791005, + "learning_rate": 6.545625630717783e-06, + "loss": 0.0001, + "step": 49270 + }, + { + "epoch": 74.10526315789474, + "grad_norm": 0.0163788590580225, + "learning_rate": 6.368281803355691e-06, + "loss": 0.0001, + "step": 49280 + }, + { + "epoch": 74.1203007518797, + "grad_norm": 0.0036826282739639282, + "learning_rate": 6.1933425226941566e-06, + "loss": 0.0002, + "step": 49290 + }, + { + "epoch": 74.13533834586467, + "grad_norm": 0.021542353555560112, + "learning_rate": 6.020809515313141e-06, + "loss": 0.0001, + "step": 49300 + }, + { + "epoch": 74.15037593984962, + "grad_norm": 0.0017033161129802465, + "learning_rate": 5.850684484043856e-06, + "loss": 0.0002, + "step": 49310 + }, + { + "epoch": 74.16541353383458, + "grad_norm": 0.0041825599037110806, + "learning_rate": 5.68296910795163e-06, + "loss": 0.0002, + "step": 49320 + }, + { + "epoch": 74.18045112781955, + "grad_norm": 0.015049166046082973, + "learning_rate": 5.517665042319542e-06, + "loss": 0.0001, + "step": 49330 + }, + { + "epoch": 74.19548872180451, + "grad_norm": 0.01032035518437624, + "learning_rate": 5.3547739186319836e-06, + "loss": 0.0001, + "step": 49340 + }, + { + "epoch": 74.21052631578948, + "grad_norm": 0.008009030483663082, + "learning_rate": 5.194297344558535e-06, + "loss": 0.0002, + "step": 49350 + }, + { + "epoch": 74.22556390977444, + "grad_norm": 0.003269106149673462, + "learning_rate": 5.0362369039382845e-06, + "loss": 0.0003, + "step": 49360 + }, + { + "epoch": 74.2406015037594, + "grad_norm": 0.002402613637968898, + "learning_rate": 4.880594156763896e-06, + "loss": 0.0001, + "step": 49370 + }, + { + "epoch": 74.25563909774436, + "grad_norm": 0.008848614990711212, + "learning_rate": 4.727370639166506e-06, + "loss": 0.0001, + "step": 49380 + }, + { + "epoch": 74.27067669172932, + "grad_norm": 0.005471036769449711, + "learning_rate": 4.5765678634003e-06, + "loss": 0.0002, + "step": 49390 + }, + { + "epoch": 74.28571428571429, + "grad_norm": 0.0012046025367453694, + "learning_rate": 4.4281873178278475e-06, + "loss": 0.0001, + "step": 49400 + }, + { + "epoch": 74.30075187969925, + "grad_norm": 0.0028437490109354258, + "learning_rate": 4.282230466905207e-06, + "loss": 0.0002, + "step": 49410 + }, + { + "epoch": 74.3157894736842, + "grad_norm": 0.011668065562844276, + "learning_rate": 4.138698751167597e-06, + "loss": 0.0001, + "step": 49420 + }, + { + "epoch": 74.33082706766918, + "grad_norm": 0.0026967115700244904, + "learning_rate": 3.997593587215076e-06, + "loss": 0.0001, + "step": 49430 + }, + { + "epoch": 74.34586466165413, + "grad_norm": 0.008883580565452576, + "learning_rate": 3.858916367698667e-06, + "loss": 0.0001, + "step": 49440 + }, + { + "epoch": 74.3609022556391, + "grad_norm": 0.006811060477048159, + "learning_rate": 3.722668461306533e-06, + "loss": 0.0001, + "step": 49450 + }, + { + "epoch": 74.37593984962406, + "grad_norm": 0.0026758676394820213, + "learning_rate": 3.588851212750488e-06, + "loss": 0.0001, + "step": 49460 + }, + { + "epoch": 74.39097744360902, + "grad_norm": 0.002077792538329959, + "learning_rate": 3.457465942752813e-06, + "loss": 0.0002, + "step": 49470 + }, + { + "epoch": 74.40601503759399, + "grad_norm": 0.007842647843062878, + "learning_rate": 3.328513948032991e-06, + "loss": 0.0001, + "step": 49480 + }, + { + "epoch": 74.42105263157895, + "grad_norm": 0.004675904754549265, + "learning_rate": 3.2019965012952125e-06, + "loss": 0.0001, + "step": 49490 + }, + { + "epoch": 74.43609022556392, + "grad_norm": 0.005686111748218536, + "learning_rate": 3.077914851215585e-06, + "loss": 0.0001, + "step": 49500 + }, + { + "epoch": 74.45112781954887, + "grad_norm": 0.005660644266754389, + "learning_rate": 2.956270222429891e-06, + "loss": 0.0002, + "step": 49510 + }, + { + "epoch": 74.46616541353383, + "grad_norm": 0.0009953785920515656, + "learning_rate": 2.8370638155215123e-06, + "loss": 0.0001, + "step": 49520 + }, + { + "epoch": 74.4812030075188, + "grad_norm": 0.013362692669034004, + "learning_rate": 2.7202968070095537e-06, + "loss": 0.0002, + "step": 49530 + }, + { + "epoch": 74.49624060150376, + "grad_norm": 0.017023183405399323, + "learning_rate": 2.6059703493372665e-06, + "loss": 0.0002, + "step": 49540 + }, + { + "epoch": 74.51127819548873, + "grad_norm": 0.0269369725137949, + "learning_rate": 2.494085570860616e-06, + "loss": 0.0002, + "step": 49550 + }, + { + "epoch": 74.52631578947368, + "grad_norm": 0.013148332014679909, + "learning_rate": 2.3846435758372033e-06, + "loss": 0.0002, + "step": 49560 + }, + { + "epoch": 74.54135338345864, + "grad_norm": 0.0016567102866247296, + "learning_rate": 2.2776454444153326e-06, + "loss": 0.0001, + "step": 49570 + }, + { + "epoch": 74.55639097744361, + "grad_norm": 0.0012478114804252982, + "learning_rate": 2.1730922326233804e-06, + "loss": 0.0001, + "step": 49580 + }, + { + "epoch": 74.57142857142857, + "grad_norm": 0.002542384434491396, + "learning_rate": 2.0709849723593023e-06, + "loss": 0.0002, + "step": 49590 + }, + { + "epoch": 74.58646616541354, + "grad_norm": 0.0019787431228905916, + "learning_rate": 1.9713246713805587e-06, + "loss": 0.0002, + "step": 49600 + }, + { + "epoch": 74.6015037593985, + "grad_norm": 0.00484830979257822, + "learning_rate": 1.8741123132940685e-06, + "loss": 0.0001, + "step": 49610 + }, + { + "epoch": 74.61654135338345, + "grad_norm": 0.0011221661698073149, + "learning_rate": 1.7793488575466032e-06, + "loss": 0.0001, + "step": 49620 + }, + { + "epoch": 74.63157894736842, + "grad_norm": 0.005124817602336407, + "learning_rate": 1.6870352394151579e-06, + "loss": 0.0001, + "step": 49630 + }, + { + "epoch": 74.64661654135338, + "grad_norm": 0.0066727264784276485, + "learning_rate": 1.5971723699979013e-06, + "loss": 0.0001, + "step": 49640 + }, + { + "epoch": 74.66165413533835, + "grad_norm": 0.0012351087061688304, + "learning_rate": 1.5097611362051012e-06, + "loss": 0.0001, + "step": 49650 + }, + { + "epoch": 74.67669172932331, + "grad_norm": 0.009319889359176159, + "learning_rate": 1.424802400750269e-06, + "loss": 0.0001, + "step": 49660 + }, + { + "epoch": 74.69172932330827, + "grad_norm": 0.007886809296905994, + "learning_rate": 1.3422970021419178e-06, + "loss": 0.0001, + "step": 49670 + }, + { + "epoch": 74.70676691729324, + "grad_norm": 0.006784120108932257, + "learning_rate": 1.2622457546749566e-06, + "loss": 0.0001, + "step": 49680 + }, + { + "epoch": 74.7218045112782, + "grad_norm": 0.014747458510100842, + "learning_rate": 1.1846494484229198e-06, + "loss": 0.0002, + "step": 49690 + }, + { + "epoch": 74.73684210526316, + "grad_norm": 0.011457344517111778, + "learning_rate": 1.109508849230001e-06, + "loss": 0.0001, + "step": 49700 + }, + { + "epoch": 74.75187969924812, + "grad_norm": 0.007363134063780308, + "learning_rate": 1.0368246987035868e-06, + "loss": 0.0002, + "step": 49710 + }, + { + "epoch": 74.76691729323308, + "grad_norm": 0.000666764157358557, + "learning_rate": 9.665977142068738e-07, + "loss": 0.0002, + "step": 49720 + }, + { + "epoch": 74.78195488721805, + "grad_norm": 0.00015817031089682132, + "learning_rate": 8.988285888519021e-07, + "loss": 0.0002, + "step": 49730 + }, + { + "epoch": 74.796992481203, + "grad_norm": 0.009564820677042007, + "learning_rate": 8.335179914925328e-07, + "loss": 0.0001, + "step": 49740 + }, + { + "epoch": 74.81203007518798, + "grad_norm": 0.0011656471760943532, + "learning_rate": 7.70666566718009e-07, + "loss": 0.0002, + "step": 49750 + }, + { + "epoch": 74.82706766917293, + "grad_norm": 0.003159643616527319, + "learning_rate": 7.102749348465165e-07, + "loss": 0.0002, + "step": 49760 + }, + { + "epoch": 74.84210526315789, + "grad_norm": 0.0019418355077505112, + "learning_rate": 6.523436919190773e-07, + "loss": 0.0001, + "step": 49770 + }, + { + "epoch": 74.85714285714286, + "grad_norm": 0.00904966238886118, + "learning_rate": 5.968734096936935e-07, + "loss": 0.0001, + "step": 49780 + }, + { + "epoch": 74.87218045112782, + "grad_norm": 0.0011590607464313507, + "learning_rate": 5.438646356396293e-07, + "loss": 0.0001, + "step": 49790 + }, + { + "epoch": 74.88721804511279, + "grad_norm": 0.012050253339111805, + "learning_rate": 4.933178929321103e-07, + "loss": 0.0001, + "step": 49800 + }, + { + "epoch": 74.90225563909775, + "grad_norm": 0.004303509835153818, + "learning_rate": 4.4523368044704915e-07, + "loss": 0.0001, + "step": 49810 + }, + { + "epoch": 74.9172932330827, + "grad_norm": 0.0026986179873347282, + "learning_rate": 3.9961247275624445e-07, + "loss": 0.0001, + "step": 49820 + }, + { + "epoch": 74.93233082706767, + "grad_norm": 0.012891637161374092, + "learning_rate": 3.5645472012257876e-07, + "loss": 0.0002, + "step": 49830 + }, + { + "epoch": 74.94736842105263, + "grad_norm": 0.0029847975820302963, + "learning_rate": 3.1576084849563315e-07, + "loss": 0.0002, + "step": 49840 + }, + { + "epoch": 74.9624060150376, + "grad_norm": 0.003847175743430853, + "learning_rate": 2.7753125950752413e-07, + "loss": 0.0001, + "step": 49850 + }, + { + "epoch": 74.97744360902256, + "grad_norm": 0.0026462466921657324, + "learning_rate": 2.4176633046882337e-07, + "loss": 0.0001, + "step": 49860 + }, + { + "epoch": 74.99248120300751, + "grad_norm": 0.004605366848409176, + "learning_rate": 2.0846641436497726e-07, + "loss": 0.0002, + "step": 49870 + }, + { + "epoch": 75.00751879699249, + "grad_norm": 0.009216357953846455, + "learning_rate": 1.7763183985269881e-07, + "loss": 0.0002, + "step": 49880 + }, + { + "epoch": 75.02255639097744, + "grad_norm": 0.00235194549895823, + "learning_rate": 1.492629112567756e-07, + "loss": 0.0002, + "step": 49890 + }, + { + "epoch": 75.0375939849624, + "grad_norm": 0.011864738538861275, + "learning_rate": 1.233599085671e-07, + "loss": 0.0002, + "step": 49900 + }, + { + "epoch": 75.05263157894737, + "grad_norm": 0.012783261947333813, + "learning_rate": 9.992308743586587e-08, + "loss": 0.0001, + "step": 49910 + }, + { + "epoch": 75.06766917293233, + "grad_norm": 0.0168161578476429, + "learning_rate": 7.895267917501503e-08, + "loss": 0.0002, + "step": 49920 + }, + { + "epoch": 75.0827067669173, + "grad_norm": 0.0064720469526946545, + "learning_rate": 6.044889075398908e-08, + "loss": 0.0002, + "step": 49930 + }, + { + "epoch": 75.09774436090225, + "grad_norm": 0.011054598726332188, + "learning_rate": 4.4411904797758695e-08, + "loss": 0.0001, + "step": 49940 + }, + { + "epoch": 75.11278195488721, + "grad_norm": 0.030894266441464424, + "learning_rate": 3.084187958485307e-08, + "loss": 0.0002, + "step": 49950 + }, + { + "epoch": 75.12781954887218, + "grad_norm": 0.022845642641186714, + "learning_rate": 1.9738949045972064e-08, + "loss": 0.0002, + "step": 49960 + }, + { + "epoch": 75.14285714285714, + "grad_norm": 0.007047316059470177, + "learning_rate": 1.1103222762542941e-08, + "loss": 0.0001, + "step": 49970 + }, + { + "epoch": 75.15789473684211, + "grad_norm": 0.018147258087992668, + "learning_rate": 4.934785965721167e-09, + "loss": 0.0002, + "step": 49980 + }, + { + "epoch": 75.17293233082707, + "grad_norm": 0.016322795301675797, + "learning_rate": 1.2336995354467196e-09, + "loss": 0.0002, + "step": 49990 + }, + { + "epoch": 75.18796992481202, + "grad_norm": 0.0027657838072627783, + "learning_rate": 0.0, + "loss": 0.0002, + "step": 50000 + }, + { + "epoch": 75.18796992481202, + "eval_cer": 0.01764392258843723, + "eval_loss": 0.12437517940998077, + "eval_runtime": 163.1943, + "eval_samples_per_second": 98.208, + "eval_steps_per_second": 0.772, + "eval_wer": 0.06509237159127215, + "step": 50000 + }, + { + "epoch": 75.18796992481202, + "step": 50000, + "total_flos": 1.4694997534629888e+18, + "train_loss": 0.12963482991572936, + "train_runtime": 27519.1016, + "train_samples_per_second": 232.566, + "train_steps_per_second": 1.817 + } + ], + "logging_steps": 10, + "max_steps": 50000, + "num_input_tokens_seen": 0, + "num_train_epochs": 76, + "save_steps": 1000, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 1.4694997534629888e+18, + "train_batch_size": 128, + "trial_name": null, + "trial_params": null +}