| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "eval_steps": 330, | |
| "global_step": 9870, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 0.0003, | |
| "loss": 33.5779, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 0.00029537037037037037, | |
| "loss": 6.0153, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_cer": 0.9522337427844499, | |
| "eval_loss": 5.343827724456787, | |
| "eval_runtime": 251.7628, | |
| "eval_samples_per_second": 10.573, | |
| "eval_steps_per_second": 0.663, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 0.0002907407407407407, | |
| "loss": 5.513, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 0.0002861111111111111, | |
| "loss": 5.3776, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.9409077230323093, | |
| "eval_loss": 5.153414726257324, | |
| "eval_runtime": 250.9411, | |
| "eval_samples_per_second": 10.608, | |
| "eval_steps_per_second": 0.665, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 0.00028148148148148146, | |
| "loss": 5.3202, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 1.37, | |
| "learning_rate": 0.0002768518518518518, | |
| "loss": 5.2604, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_cer": 0.9108417913714333, | |
| "eval_loss": 5.083229064941406, | |
| "eval_runtime": 252.1577, | |
| "eval_samples_per_second": 10.557, | |
| "eval_steps_per_second": 0.662, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "learning_rate": 0.0002722222222222222, | |
| "loss": 5.2287, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 0.00026759259259259255, | |
| "loss": 5.2393, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "eval_cer": 0.9073126692747517, | |
| "eval_loss": 5.065478324890137, | |
| "eval_runtime": 251.9951, | |
| "eval_samples_per_second": 10.564, | |
| "eval_steps_per_second": 0.663, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 0.00026296296296296294, | |
| "loss": 5.2058, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 2.28, | |
| "learning_rate": 0.00025833333333333334, | |
| "loss": 5.1796, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "learning_rate": 0.0002537037037037037, | |
| "loss": 5.1721, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 2.51, | |
| "eval_cer": 0.9000355647963232, | |
| "eval_loss": 5.046383380889893, | |
| "eval_runtime": 251.3928, | |
| "eval_samples_per_second": 10.589, | |
| "eval_steps_per_second": 0.664, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "learning_rate": 0.0002490740740740741, | |
| "loss": 5.1736, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 0.00024444444444444443, | |
| "loss": 5.1619, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 3.01, | |
| "eval_cer": 0.9044674855689, | |
| "eval_loss": 5.024378776550293, | |
| "eval_runtime": 248.9825, | |
| "eval_samples_per_second": 10.692, | |
| "eval_steps_per_second": 0.671, | |
| "step": 1980 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 0.0002398148148148148, | |
| "loss": 5.1484, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "learning_rate": 0.00023518518518518517, | |
| "loss": 5.1308, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "eval_cer": 0.9020326648975461, | |
| "eval_loss": 5.021634578704834, | |
| "eval_runtime": 250.5536, | |
| "eval_samples_per_second": 10.624, | |
| "eval_steps_per_second": 0.667, | |
| "step": 2310 | |
| }, | |
| { | |
| "epoch": 3.65, | |
| "learning_rate": 0.00023055555555555552, | |
| "loss": 5.0855, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.88, | |
| "learning_rate": 0.00022592592592592591, | |
| "loss": 5.0971, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 4.01, | |
| "eval_cer": 0.9040297649987689, | |
| "eval_loss": 4.9340667724609375, | |
| "eval_runtime": 248.8115, | |
| "eval_samples_per_second": 10.699, | |
| "eval_steps_per_second": 0.671, | |
| "step": 2640 | |
| }, | |
| { | |
| "epoch": 4.1, | |
| "learning_rate": 0.00022129629629629626, | |
| "loss": 5.0599, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 0.00021666666666666666, | |
| "loss": 5.0137, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 4.51, | |
| "eval_cer": 0.9143709134681148, | |
| "eval_loss": 4.879497051239014, | |
| "eval_runtime": 248.574, | |
| "eval_samples_per_second": 10.709, | |
| "eval_steps_per_second": 0.672, | |
| "step": 2970 | |
| }, | |
| { | |
| "epoch": 4.56, | |
| "learning_rate": 0.00021203703703703703, | |
| "loss": 4.9809, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.79, | |
| "learning_rate": 0.00020740740740740737, | |
| "loss": 4.9939, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "learning_rate": 0.00020277777777777777, | |
| "loss": 4.9341, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 5.02, | |
| "eval_cer": 0.9039476923918693, | |
| "eval_loss": 4.725036144256592, | |
| "eval_runtime": 250.9019, | |
| "eval_samples_per_second": 10.61, | |
| "eval_steps_per_second": 0.666, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "learning_rate": 0.00019814814814814814, | |
| "loss": 4.8114, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 5.47, | |
| "learning_rate": 0.00019351851851851849, | |
| "loss": 4.6832, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 5.52, | |
| "eval_cer": 0.8367302273411211, | |
| "eval_loss": 4.214047908782959, | |
| "eval_runtime": 249.955, | |
| "eval_samples_per_second": 10.65, | |
| "eval_steps_per_second": 0.668, | |
| "step": 3630 | |
| }, | |
| { | |
| "epoch": 5.7, | |
| "learning_rate": 0.00018888888888888888, | |
| "loss": 4.4588, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 5.93, | |
| "learning_rate": 0.00018425925925925923, | |
| "loss": 4.1627, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "eval_cer": 0.7318140781878368, | |
| "eval_loss": 3.4010486602783203, | |
| "eval_runtime": 249.3206, | |
| "eval_samples_per_second": 10.677, | |
| "eval_steps_per_second": 0.67, | |
| "step": 3960 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "learning_rate": 0.0001796296296296296, | |
| "loss": 3.7597, | |
| "step": 4050 | |
| }, | |
| { | |
| "epoch": 6.38, | |
| "learning_rate": 0.000175, | |
| "loss": 3.5448, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "eval_cer": 0.6479905890077422, | |
| "eval_loss": 2.882997989654541, | |
| "eval_runtime": 248.069, | |
| "eval_samples_per_second": 10.731, | |
| "eval_steps_per_second": 0.673, | |
| "step": 4290 | |
| }, | |
| { | |
| "epoch": 6.61, | |
| "learning_rate": 0.00017037037037037034, | |
| "loss": 3.3922, | |
| "step": 4350 | |
| }, | |
| { | |
| "epoch": 6.84, | |
| "learning_rate": 0.00016574074074074074, | |
| "loss": 3.2576, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 7.02, | |
| "eval_cer": 0.6265696386069542, | |
| "eval_loss": 2.6253392696380615, | |
| "eval_runtime": 244.3615, | |
| "eval_samples_per_second": 10.894, | |
| "eval_steps_per_second": 0.683, | |
| "step": 4620 | |
| }, | |
| { | |
| "epoch": 7.07, | |
| "learning_rate": 0.0001611111111111111, | |
| "loss": 3.0846, | |
| "step": 4650 | |
| }, | |
| { | |
| "epoch": 7.29, | |
| "learning_rate": 0.00015648148148148146, | |
| "loss": 2.9344, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 7.52, | |
| "learning_rate": 0.00015185185185185185, | |
| "loss": 2.8561, | |
| "step": 4950 | |
| }, | |
| { | |
| "epoch": 7.52, | |
| "eval_cer": 0.5866002790468634, | |
| "eval_loss": 2.430042266845703, | |
| "eval_runtime": 245.691, | |
| "eval_samples_per_second": 10.835, | |
| "eval_steps_per_second": 0.68, | |
| "step": 4950 | |
| }, | |
| { | |
| "epoch": 7.75, | |
| "learning_rate": 0.00014722222222222223, | |
| "loss": 2.8167, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 7.98, | |
| "learning_rate": 0.00014259259259259257, | |
| "loss": 2.7894, | |
| "step": 5250 | |
| }, | |
| { | |
| "epoch": 8.02, | |
| "eval_cer": 0.575028041474024, | |
| "eval_loss": 2.2997841835021973, | |
| "eval_runtime": 245.9682, | |
| "eval_samples_per_second": 10.823, | |
| "eval_steps_per_second": 0.679, | |
| "step": 5280 | |
| }, | |
| { | |
| "epoch": 8.21, | |
| "learning_rate": 0.00013796296296296294, | |
| "loss": 2.6472, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 8.43, | |
| "learning_rate": 0.0001333333333333333, | |
| "loss": 2.6018, | |
| "step": 5550 | |
| }, | |
| { | |
| "epoch": 8.53, | |
| "eval_cer": 0.554892895247996, | |
| "eval_loss": 2.187838554382324, | |
| "eval_runtime": 244.7687, | |
| "eval_samples_per_second": 10.876, | |
| "eval_steps_per_second": 0.682, | |
| "step": 5610 | |
| }, | |
| { | |
| "epoch": 8.66, | |
| "learning_rate": 0.0001287037037037037, | |
| "loss": 2.5751, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 8.89, | |
| "learning_rate": 0.00012407407407407406, | |
| "loss": 2.546, | |
| "step": 5850 | |
| }, | |
| { | |
| "epoch": 9.03, | |
| "eval_cer": 0.5350860394495663, | |
| "eval_loss": 2.1450469493865967, | |
| "eval_runtime": 244.8384, | |
| "eval_samples_per_second": 10.872, | |
| "eval_steps_per_second": 0.682, | |
| "step": 5940 | |
| }, | |
| { | |
| "epoch": 9.12, | |
| "learning_rate": 0.00011944444444444443, | |
| "loss": 2.4555, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 9.35, | |
| "learning_rate": 0.0001148148148148148, | |
| "loss": 2.3787, | |
| "step": 6150 | |
| }, | |
| { | |
| "epoch": 9.53, | |
| "eval_cer": 0.5339917380242388, | |
| "eval_loss": 2.102729558944702, | |
| "eval_runtime": 244.3347, | |
| "eval_samples_per_second": 10.895, | |
| "eval_steps_per_second": 0.683, | |
| "step": 6270 | |
| }, | |
| { | |
| "epoch": 9.57, | |
| "learning_rate": 0.00011018518518518518, | |
| "loss": 2.3783, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 9.8, | |
| "learning_rate": 0.00010555555555555555, | |
| "loss": 2.3806, | |
| "step": 6450 | |
| }, | |
| { | |
| "epoch": 10.03, | |
| "learning_rate": 0.00010092592592592591, | |
| "loss": 2.335, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 10.03, | |
| "eval_cer": 0.5165923453615299, | |
| "eval_loss": 2.0303709506988525, | |
| "eval_runtime": 244.6547, | |
| "eval_samples_per_second": 10.881, | |
| "eval_steps_per_second": 0.683, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 10.26, | |
| "learning_rate": 9.629629629629628e-05, | |
| "loss": 2.2373, | |
| "step": 6750 | |
| }, | |
| { | |
| "epoch": 10.49, | |
| "learning_rate": 9.166666666666667e-05, | |
| "loss": 2.2138, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 10.53, | |
| "eval_cer": 0.5164555576833639, | |
| "eval_loss": 2.0100014209747314, | |
| "eval_runtime": 244.9943, | |
| "eval_samples_per_second": 10.866, | |
| "eval_steps_per_second": 0.682, | |
| "step": 6930 | |
| }, | |
| { | |
| "epoch": 10.71, | |
| "learning_rate": 8.703703703703704e-05, | |
| "loss": 2.2174, | |
| "step": 7050 | |
| }, | |
| { | |
| "epoch": 10.94, | |
| "learning_rate": 8.24074074074074e-05, | |
| "loss": 2.2381, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 11.03, | |
| "eval_cer": 0.5031050802943671, | |
| "eval_loss": 1.9650695323944092, | |
| "eval_runtime": 244.4808, | |
| "eval_samples_per_second": 10.888, | |
| "eval_steps_per_second": 0.683, | |
| "step": 7260 | |
| }, | |
| { | |
| "epoch": 11.17, | |
| "learning_rate": 7.777777777777777e-05, | |
| "loss": 2.155, | |
| "step": 7350 | |
| }, | |
| { | |
| "epoch": 11.4, | |
| "learning_rate": 7.314814814814814e-05, | |
| "loss": 2.1108, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 11.53, | |
| "eval_cer": 0.5034880857932318, | |
| "eval_loss": 1.9666314125061035, | |
| "eval_runtime": 244.872, | |
| "eval_samples_per_second": 10.871, | |
| "eval_steps_per_second": 0.682, | |
| "step": 7590 | |
| }, | |
| { | |
| "epoch": 11.63, | |
| "learning_rate": 6.851851851851851e-05, | |
| "loss": 2.1221, | |
| "step": 7650 | |
| }, | |
| { | |
| "epoch": 11.85, | |
| "learning_rate": 6.388888888888888e-05, | |
| "loss": 2.0916, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 12.04, | |
| "eval_cer": 0.49982217601838425, | |
| "eval_loss": 1.913594365119934, | |
| "eval_runtime": 244.907, | |
| "eval_samples_per_second": 10.869, | |
| "eval_steps_per_second": 0.682, | |
| "step": 7920 | |
| }, | |
| { | |
| "epoch": 12.08, | |
| "learning_rate": 5.925925925925925e-05, | |
| "loss": 2.0517, | |
| "step": 7950 | |
| }, | |
| { | |
| "epoch": 12.31, | |
| "learning_rate": 5.4629629629629624e-05, | |
| "loss": 2.0279, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 12.54, | |
| "learning_rate": 4.9999999999999996e-05, | |
| "loss": 2.0229, | |
| "step": 8250 | |
| }, | |
| { | |
| "epoch": 12.54, | |
| "eval_cer": 0.5027767898667688, | |
| "eval_loss": 1.898772120475769, | |
| "eval_runtime": 244.3507, | |
| "eval_samples_per_second": 10.894, | |
| "eval_steps_per_second": 0.683, | |
| "step": 8250 | |
| }, | |
| { | |
| "epoch": 12.77, | |
| "learning_rate": 4.537037037037037e-05, | |
| "loss": 1.9891, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 12.99, | |
| "learning_rate": 4.074074074074074e-05, | |
| "loss": 2.0056, | |
| "step": 8550 | |
| }, | |
| { | |
| "epoch": 13.04, | |
| "eval_cer": 0.4995759581976856, | |
| "eval_loss": 1.8768519163131714, | |
| "eval_runtime": 244.2609, | |
| "eval_samples_per_second": 10.898, | |
| "eval_steps_per_second": 0.684, | |
| "step": 8580 | |
| }, | |
| { | |
| "epoch": 13.22, | |
| "learning_rate": 3.61111111111111e-05, | |
| "loss": 1.9451, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 13.45, | |
| "learning_rate": 3.1481481481481474e-05, | |
| "loss": 1.9245, | |
| "step": 8850 | |
| }, | |
| { | |
| "epoch": 13.54, | |
| "eval_cer": 0.495472327852707, | |
| "eval_loss": 1.8715523481369019, | |
| "eval_runtime": 244.453, | |
| "eval_samples_per_second": 10.89, | |
| "eval_steps_per_second": 0.683, | |
| "step": 8910 | |
| }, | |
| { | |
| "epoch": 13.68, | |
| "learning_rate": 2.685185185185185e-05, | |
| "loss": 1.916, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 13.91, | |
| "learning_rate": 2.222222222222222e-05, | |
| "loss": 1.9378, | |
| "step": 9150 | |
| }, | |
| { | |
| "epoch": 14.04, | |
| "eval_cer": 0.49459688671244495, | |
| "eval_loss": 1.8560909032821655, | |
| "eval_runtime": 244.76, | |
| "eval_samples_per_second": 10.876, | |
| "eval_steps_per_second": 0.682, | |
| "step": 9240 | |
| }, | |
| { | |
| "epoch": 14.13, | |
| "learning_rate": 1.759259259259259e-05, | |
| "loss": 1.9169, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 14.36, | |
| "learning_rate": 1.296296296296296e-05, | |
| "loss": 1.9003, | |
| "step": 9450 | |
| }, | |
| { | |
| "epoch": 14.54, | |
| "eval_cer": 0.49363937296528326, | |
| "eval_loss": 1.848546028137207, | |
| "eval_runtime": 244.6555, | |
| "eval_samples_per_second": 10.881, | |
| "eval_steps_per_second": 0.683, | |
| "step": 9570 | |
| }, | |
| { | |
| "epoch": 14.59, | |
| "learning_rate": 8.333333333333332e-06, | |
| "loss": 1.8611, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 14.82, | |
| "learning_rate": 3.7037037037037033e-06, | |
| "loss": 1.8698, | |
| "step": 9750 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "step": 9870, | |
| "total_flos": 3.0352534937393955e+19, | |
| "train_loss": 3.8899900426739014, | |
| "train_runtime": 13723.3522, | |
| "train_samples_per_second": 23.006, | |
| "train_steps_per_second": 0.719 | |
| } | |
| ], | |
| "logging_steps": 150, | |
| "max_steps": 9870, | |
| "num_train_epochs": 15, | |
| "save_steps": 330, | |
| "total_flos": 3.0352534937393955e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |