| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 13.465952563121652, | |
| "eval_steps": 2000, | |
| "global_step": 22000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 3.9999999999999996e-05, | |
| "loss": 33.0181, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 7.999999999999999e-05, | |
| "loss": 8.7646, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 0.00011999999999999999, | |
| "loss": 4.6933, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 0.00015999999999999999, | |
| "loss": 4.5796, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 0.00019999999999999998, | |
| "loss": 4.1047, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 0.00023999999999999998, | |
| "loss": 2.6327, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 0.00028, | |
| "loss": 1.9893, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 0.00029869536855838224, | |
| "loss": 1.7241, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 0.00029608610567514673, | |
| "loss": 1.5607, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "learning_rate": 0.00029347684279191127, | |
| "loss": 1.4503, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.22, | |
| "eval_cer": 0.268688843618788, | |
| "eval_loss": 1.0610458850860596, | |
| "eval_runtime": 1586.9419, | |
| "eval_samples_per_second": 8.25, | |
| "eval_steps_per_second": 0.516, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 0.00029086757990867576, | |
| "loss": 1.3782, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.47, | |
| "learning_rate": 0.0002882583170254403, | |
| "loss": 1.3244, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "learning_rate": 0.0002856490541422048, | |
| "loss": 1.2553, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 0.00028303979125896933, | |
| "loss": 1.2077, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 0.0002804305283757338, | |
| "loss": 1.1868, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.96, | |
| "learning_rate": 0.00027782126549249836, | |
| "loss": 1.1624, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 0.00027521200260926284, | |
| "loss": 1.1136, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 0.0002726027397260274, | |
| "loss": 1.077, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 0.00026999347684279187, | |
| "loss": 1.053, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 0.0002673842139595564, | |
| "loss": 1.0239, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "eval_cer": 0.19039349567825709, | |
| "eval_loss": 0.6961866617202759, | |
| "eval_runtime": 1607.479, | |
| "eval_samples_per_second": 8.145, | |
| "eval_steps_per_second": 0.509, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 0.0002647749510763209, | |
| "loss": 1.0206, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "learning_rate": 0.00026216568819308544, | |
| "loss": 1.0045, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "learning_rate": 0.0002595564253098499, | |
| "loss": 0.9802, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 2.94, | |
| "learning_rate": 0.00025694716242661447, | |
| "loss": 0.9839, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 3.06, | |
| "learning_rate": 0.00025433789954337895, | |
| "loss": 0.9243, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.18, | |
| "learning_rate": 0.0002517286366601435, | |
| "loss": 0.9082, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 3.31, | |
| "learning_rate": 0.000249119373776908, | |
| "loss": 0.9017, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 0.0002465101108936725, | |
| "loss": 0.8871, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 3.55, | |
| "learning_rate": 0.00024390084801043704, | |
| "loss": 0.9036, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 0.00024129158512720155, | |
| "loss": 0.8977, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "eval_cer": 0.16872254319465907, | |
| "eval_loss": 0.594495415687561, | |
| "eval_runtime": 1591.6049, | |
| "eval_samples_per_second": 8.226, | |
| "eval_steps_per_second": 0.515, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "learning_rate": 0.00023868232224396607, | |
| "loss": 0.8843, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "learning_rate": 0.00023607305936073058, | |
| "loss": 0.8757, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 0.0002334637964774951, | |
| "loss": 0.8573, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 4.16, | |
| "learning_rate": 0.0002308545335942596, | |
| "loss": 0.8126, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 4.28, | |
| "learning_rate": 0.00022824527071102412, | |
| "loss": 0.8192, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 4.41, | |
| "learning_rate": 0.00022563600782778863, | |
| "loss": 0.8061, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 4.53, | |
| "learning_rate": 0.00022302674494455315, | |
| "loss": 0.8123, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 4.65, | |
| "learning_rate": 0.00022041748206131766, | |
| "loss": 0.8046, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 4.77, | |
| "learning_rate": 0.00021780821917808218, | |
| "loss": 0.7979, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 4.9, | |
| "learning_rate": 0.0002151989562948467, | |
| "loss": 0.804, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 4.9, | |
| "eval_cer": 0.14924661713942214, | |
| "eval_loss": 0.5327703952789307, | |
| "eval_runtime": 1595.6324, | |
| "eval_samples_per_second": 8.206, | |
| "eval_steps_per_second": 0.513, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "learning_rate": 0.0002125896934116112, | |
| "loss": 0.7867, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "learning_rate": 0.00020998043052837572, | |
| "loss": 0.7557, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 5.26, | |
| "learning_rate": 0.00020737116764514023, | |
| "loss": 0.7478, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 5.39, | |
| "learning_rate": 0.00020476190476190475, | |
| "loss": 0.7398, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 5.51, | |
| "learning_rate": 0.00020215264187866926, | |
| "loss": 0.7408, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 5.63, | |
| "learning_rate": 0.00019954337899543377, | |
| "loss": 0.75, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "learning_rate": 0.0001969341161121983, | |
| "loss": 0.7344, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 5.88, | |
| "learning_rate": 0.0001943248532289628, | |
| "loss": 0.738, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 0.00019171559034572732, | |
| "loss": 0.7373, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 6.12, | |
| "learning_rate": 0.00018910632746249183, | |
| "loss": 0.698, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 6.12, | |
| "eval_cer": 0.13653489424101573, | |
| "eval_loss": 0.5013594031333923, | |
| "eval_runtime": 1605.1161, | |
| "eval_samples_per_second": 8.157, | |
| "eval_steps_per_second": 0.51, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 6.24, | |
| "learning_rate": 0.00018649706457925634, | |
| "loss": 0.6943, | |
| "step": 10200 | |
| }, | |
| { | |
| "epoch": 6.37, | |
| "learning_rate": 0.00018388780169602086, | |
| "loss": 0.6997, | |
| "step": 10400 | |
| }, | |
| { | |
| "epoch": 6.49, | |
| "learning_rate": 0.00018127853881278537, | |
| "loss": 0.6929, | |
| "step": 10600 | |
| }, | |
| { | |
| "epoch": 6.61, | |
| "learning_rate": 0.00017866927592954989, | |
| "loss": 0.7003, | |
| "step": 10800 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "learning_rate": 0.0001760600130463144, | |
| "loss": 0.6863, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 6.86, | |
| "learning_rate": 0.00017345075016307891, | |
| "loss": 0.6883, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 6.98, | |
| "learning_rate": 0.00017084148727984343, | |
| "loss": 0.6787, | |
| "step": 11400 | |
| }, | |
| { | |
| "epoch": 7.1, | |
| "learning_rate": 0.00016823222439660794, | |
| "loss": 0.6518, | |
| "step": 11600 | |
| }, | |
| { | |
| "epoch": 7.22, | |
| "learning_rate": 0.00016562296151337246, | |
| "loss": 0.6494, | |
| "step": 11800 | |
| }, | |
| { | |
| "epoch": 7.35, | |
| "learning_rate": 0.00016301369863013697, | |
| "loss": 0.6426, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 7.35, | |
| "eval_cer": 0.13216305737125092, | |
| "eval_loss": 0.47150149941444397, | |
| "eval_runtime": 1597.7342, | |
| "eval_samples_per_second": 8.195, | |
| "eval_steps_per_second": 0.513, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 7.47, | |
| "learning_rate": 0.00016040443574690148, | |
| "loss": 0.6457, | |
| "step": 12200 | |
| }, | |
| { | |
| "epoch": 7.59, | |
| "learning_rate": 0.000157795172863666, | |
| "loss": 0.6429, | |
| "step": 12400 | |
| }, | |
| { | |
| "epoch": 7.71, | |
| "learning_rate": 0.0001551859099804305, | |
| "loss": 0.6512, | |
| "step": 12600 | |
| }, | |
| { | |
| "epoch": 7.83, | |
| "learning_rate": 0.00015257664709719503, | |
| "loss": 0.6458, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 7.96, | |
| "learning_rate": 0.00014996738421395954, | |
| "loss": 0.6501, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 8.08, | |
| "learning_rate": 0.00014735812133072405, | |
| "loss": 0.6095, | |
| "step": 13200 | |
| }, | |
| { | |
| "epoch": 8.2, | |
| "learning_rate": 0.00014474885844748857, | |
| "loss": 0.6132, | |
| "step": 13400 | |
| }, | |
| { | |
| "epoch": 8.32, | |
| "learning_rate": 0.00014213959556425308, | |
| "loss": 0.6102, | |
| "step": 13600 | |
| }, | |
| { | |
| "epoch": 8.45, | |
| "learning_rate": 0.0001395303326810176, | |
| "loss": 0.6125, | |
| "step": 13800 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "learning_rate": 0.0001369210697977821, | |
| "loss": 0.61, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "eval_cer": 0.1257692459492199, | |
| "eval_loss": 0.45295360684394836, | |
| "eval_runtime": 1603.0798, | |
| "eval_samples_per_second": 8.167, | |
| "eval_steps_per_second": 0.511, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 8.69, | |
| "learning_rate": 0.00013431180691454662, | |
| "loss": 0.606, | |
| "step": 14200 | |
| }, | |
| { | |
| "epoch": 8.81, | |
| "learning_rate": 0.00013170254403131114, | |
| "loss": 0.5957, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 8.94, | |
| "learning_rate": 0.00012909328114807565, | |
| "loss": 0.5992, | |
| "step": 14600 | |
| }, | |
| { | |
| "epoch": 9.06, | |
| "learning_rate": 0.00012648401826484017, | |
| "loss": 0.5752, | |
| "step": 14800 | |
| }, | |
| { | |
| "epoch": 9.18, | |
| "learning_rate": 0.00012387475538160468, | |
| "loss": 0.5654, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 9.3, | |
| "learning_rate": 0.00012126549249836919, | |
| "loss": 0.5725, | |
| "step": 15200 | |
| }, | |
| { | |
| "epoch": 9.43, | |
| "learning_rate": 0.00011865622961513371, | |
| "loss": 0.5713, | |
| "step": 15400 | |
| }, | |
| { | |
| "epoch": 9.55, | |
| "learning_rate": 0.00011604696673189822, | |
| "loss": 0.5649, | |
| "step": 15600 | |
| }, | |
| { | |
| "epoch": 9.67, | |
| "learning_rate": 0.00011343770384866273, | |
| "loss": 0.5643, | |
| "step": 15800 | |
| }, | |
| { | |
| "epoch": 9.79, | |
| "learning_rate": 0.00011082844096542725, | |
| "loss": 0.5709, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 9.79, | |
| "eval_cer": 0.1200554980402634, | |
| "eval_loss": 0.4299587607383728, | |
| "eval_runtime": 1609.5227, | |
| "eval_samples_per_second": 8.135, | |
| "eval_steps_per_second": 0.509, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 9.92, | |
| "learning_rate": 0.00010821917808219176, | |
| "loss": 0.5666, | |
| "step": 16200 | |
| }, | |
| { | |
| "epoch": 10.04, | |
| "learning_rate": 0.00010560991519895628, | |
| "loss": 0.5531, | |
| "step": 16400 | |
| }, | |
| { | |
| "epoch": 10.16, | |
| "learning_rate": 0.00010300065231572079, | |
| "loss": 0.5389, | |
| "step": 16600 | |
| }, | |
| { | |
| "epoch": 10.28, | |
| "learning_rate": 0.0001003913894324853, | |
| "loss": 0.5456, | |
| "step": 16800 | |
| }, | |
| { | |
| "epoch": 10.41, | |
| "learning_rate": 9.778212654924982e-05, | |
| "loss": 0.5353, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 10.53, | |
| "learning_rate": 9.517286366601433e-05, | |
| "loss": 0.5337, | |
| "step": 17200 | |
| }, | |
| { | |
| "epoch": 10.65, | |
| "learning_rate": 9.256360078277885e-05, | |
| "loss": 0.5296, | |
| "step": 17400 | |
| }, | |
| { | |
| "epoch": 10.77, | |
| "learning_rate": 8.995433789954336e-05, | |
| "loss": 0.5372, | |
| "step": 17600 | |
| }, | |
| { | |
| "epoch": 10.9, | |
| "learning_rate": 8.734507501630787e-05, | |
| "loss": 0.5388, | |
| "step": 17800 | |
| }, | |
| { | |
| "epoch": 11.02, | |
| "learning_rate": 8.473581213307239e-05, | |
| "loss": 0.5235, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 11.02, | |
| "eval_cer": 0.11664607248141211, | |
| "eval_loss": 0.4167773723602295, | |
| "eval_runtime": 1608.2913, | |
| "eval_samples_per_second": 8.141, | |
| "eval_steps_per_second": 0.509, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 11.14, | |
| "learning_rate": 8.212654924983692e-05, | |
| "loss": 0.509, | |
| "step": 18200 | |
| }, | |
| { | |
| "epoch": 11.26, | |
| "learning_rate": 7.951728636660143e-05, | |
| "loss": 0.5116, | |
| "step": 18400 | |
| }, | |
| { | |
| "epoch": 11.38, | |
| "learning_rate": 7.690802348336594e-05, | |
| "loss": 0.4967, | |
| "step": 18600 | |
| }, | |
| { | |
| "epoch": 11.51, | |
| "learning_rate": 7.429876060013046e-05, | |
| "loss": 0.511, | |
| "step": 18800 | |
| }, | |
| { | |
| "epoch": 11.63, | |
| "learning_rate": 7.168949771689497e-05, | |
| "loss": 0.5056, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 11.75, | |
| "learning_rate": 6.908023483365949e-05, | |
| "loss": 0.5073, | |
| "step": 19200 | |
| }, | |
| { | |
| "epoch": 11.87, | |
| "learning_rate": 6.6470971950424e-05, | |
| "loss": 0.4968, | |
| "step": 19400 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 6.386170906718851e-05, | |
| "loss": 0.5009, | |
| "step": 19600 | |
| }, | |
| { | |
| "epoch": 12.12, | |
| "learning_rate": 6.125244618395303e-05, | |
| "loss": 0.4832, | |
| "step": 19800 | |
| }, | |
| { | |
| "epoch": 12.24, | |
| "learning_rate": 5.864318330071754e-05, | |
| "loss": 0.4778, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 12.24, | |
| "eval_cer": 0.11294822712906938, | |
| "eval_loss": 0.40570223331451416, | |
| "eval_runtime": 1612.661, | |
| "eval_samples_per_second": 8.119, | |
| "eval_steps_per_second": 0.508, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 12.36, | |
| "learning_rate": 5.6033920417482055e-05, | |
| "loss": 0.4775, | |
| "step": 20200 | |
| }, | |
| { | |
| "epoch": 12.49, | |
| "learning_rate": 5.342465753424657e-05, | |
| "loss": 0.4855, | |
| "step": 20400 | |
| }, | |
| { | |
| "epoch": 12.61, | |
| "learning_rate": 5.081539465101108e-05, | |
| "loss": 0.4773, | |
| "step": 20600 | |
| }, | |
| { | |
| "epoch": 12.73, | |
| "learning_rate": 4.82061317677756e-05, | |
| "loss": 0.4745, | |
| "step": 20800 | |
| }, | |
| { | |
| "epoch": 12.85, | |
| "learning_rate": 4.559686888454011e-05, | |
| "loss": 0.48, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 12.98, | |
| "learning_rate": 4.2987606001304625e-05, | |
| "loss": 0.463, | |
| "step": 21200 | |
| }, | |
| { | |
| "epoch": 13.1, | |
| "learning_rate": 4.037834311806914e-05, | |
| "loss": 0.4643, | |
| "step": 21400 | |
| }, | |
| { | |
| "epoch": 13.22, | |
| "learning_rate": 3.776908023483365e-05, | |
| "loss": 0.449, | |
| "step": 21600 | |
| }, | |
| { | |
| "epoch": 13.34, | |
| "learning_rate": 3.515981735159817e-05, | |
| "loss": 0.4604, | |
| "step": 21800 | |
| }, | |
| { | |
| "epoch": 13.47, | |
| "learning_rate": 3.255055446836268e-05, | |
| "loss": 0.4571, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 13.47, | |
| "eval_cer": 0.10995473327241098, | |
| "eval_loss": 0.3945465385913849, | |
| "eval_runtime": 1610.0901, | |
| "eval_samples_per_second": 8.132, | |
| "eval_steps_per_second": 0.509, | |
| "step": 22000 | |
| } | |
| ], | |
| "logging_steps": 200, | |
| "max_steps": 24495, | |
| "num_train_epochs": 15, | |
| "save_steps": 2000, | |
| "total_flos": 1.5466215358461693e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |