| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 8.0, | |
| "eval_steps": 500, | |
| "global_step": 960, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 6.000000000000001e-07, | |
| "loss": 16.3378, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 1.2000000000000002e-06, | |
| "loss": 16.7146, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 1.8e-06, | |
| "loss": 16.2989, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "loss": 15.6823, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3e-06, | |
| "loss": 15.2213, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 3.6e-06, | |
| "loss": 13.8045, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 4.2000000000000004e-06, | |
| "loss": 12.3665, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 4.800000000000001e-06, | |
| "loss": 9.9596, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 5.4e-06, | |
| "loss": 7.1678, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 6e-06, | |
| "loss": 5.6617, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 6.6e-06, | |
| "loss": 4.5458, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 7.2e-06, | |
| "loss": 3.4904, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.999086184248101, | |
| "eval_loss": 3.4430177211761475, | |
| "eval_runtime": 13.8851, | |
| "eval_samples_per_second": 69.139, | |
| "eval_steps_per_second": 2.161, | |
| "eval_wer": 0.996996996996997, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.08, | |
| "learning_rate": 7.8e-06, | |
| "loss": 3.2384, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.17, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 2.8932, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.25, | |
| "learning_rate": 9e-06, | |
| "loss": 2.7276, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.33, | |
| "learning_rate": 9.600000000000001e-06, | |
| "loss": 2.6587, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 1.02e-05, | |
| "loss": 2.3907, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 1.08e-05, | |
| "loss": 2.22, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 1.1400000000000001e-05, | |
| "loss": 2.0003, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 1.2e-05, | |
| "loss": 1.8618, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.75, | |
| "learning_rate": 1.26e-05, | |
| "loss": 1.6286, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.83, | |
| "learning_rate": 1.32e-05, | |
| "loss": 1.4244, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.92, | |
| "learning_rate": 1.3800000000000002e-05, | |
| "loss": 1.3615, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 1.44e-05, | |
| "loss": 1.1939, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.6264778114112742, | |
| "eval_loss": 1.006431221961975, | |
| "eval_runtime": 15.0117, | |
| "eval_samples_per_second": 63.95, | |
| "eval_steps_per_second": 1.998, | |
| "eval_wer": 0.8269519519519519, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 1.5e-05, | |
| "loss": 1.3824, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.17, | |
| "learning_rate": 1.56e-05, | |
| "loss": 1.0939, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.25, | |
| "learning_rate": 1.62e-05, | |
| "loss": 1.0596, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.33, | |
| "learning_rate": 1.6800000000000002e-05, | |
| "loss": 1.0087, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "learning_rate": 1.74e-05, | |
| "loss": 0.9328, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 1.8e-05, | |
| "loss": 0.9045, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.58, | |
| "learning_rate": 1.86e-05, | |
| "loss": 0.8645, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 1.9200000000000003e-05, | |
| "loss": 0.8474, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.75, | |
| "learning_rate": 1.98e-05, | |
| "loss": 0.7985, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 2.04e-05, | |
| "loss": 0.7874, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.92, | |
| "learning_rate": 2.1e-05, | |
| "loss": 0.8, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 2.16e-05, | |
| "loss": 0.7726, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.5705065966074591, | |
| "eval_loss": 0.6256773471832275, | |
| "eval_runtime": 13.9464, | |
| "eval_samples_per_second": 68.835, | |
| "eval_steps_per_second": 2.151, | |
| "eval_wer": 0.8198198198198198, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 3.08, | |
| "learning_rate": 2.22e-05, | |
| "loss": 0.7963, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 3.17, | |
| "learning_rate": 2.2800000000000002e-05, | |
| "loss": 0.7342, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 3.25, | |
| "learning_rate": 2.3400000000000003e-05, | |
| "loss": 0.7324, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 2.4e-05, | |
| "loss": 0.6865, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.42, | |
| "learning_rate": 2.4599999999999998e-05, | |
| "loss": 0.6731, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "learning_rate": 2.52e-05, | |
| "loss": 0.6683, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 3.58, | |
| "learning_rate": 2.58e-05, | |
| "loss": 0.6583, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 3.67, | |
| "learning_rate": 2.64e-05, | |
| "loss": 0.6218, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "learning_rate": 2.7000000000000002e-05, | |
| "loss": 0.5825, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "learning_rate": 2.7600000000000003e-05, | |
| "loss": 0.5552, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 3.92, | |
| "learning_rate": 2.8199999999999998e-05, | |
| "loss": 0.6122, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 2.88e-05, | |
| "loss": 0.5502, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.34153863727225997, | |
| "eval_loss": 0.41475021839141846, | |
| "eval_runtime": 13.9543, | |
| "eval_samples_per_second": 68.796, | |
| "eval_steps_per_second": 2.15, | |
| "eval_wer": 0.5910285285285285, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 4.08, | |
| "learning_rate": 2.94e-05, | |
| "loss": 0.5787, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "learning_rate": 3e-05, | |
| "loss": 0.4964, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "learning_rate": 3.06e-05, | |
| "loss": 0.5245, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 4.33, | |
| "learning_rate": 3.12e-05, | |
| "loss": 0.4688, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 4.42, | |
| "learning_rate": 3.18e-05, | |
| "loss": 0.5043, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 3.24e-05, | |
| "loss": 0.4769, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 4.58, | |
| "learning_rate": 3.3e-05, | |
| "loss": 0.4966, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "learning_rate": 3.3600000000000004e-05, | |
| "loss": 0.4772, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "learning_rate": 3.42e-05, | |
| "loss": 0.4364, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 3.48e-05, | |
| "loss": 0.417, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 4.92, | |
| "learning_rate": 3.54e-05, | |
| "loss": 0.4407, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 3.6e-05, | |
| "loss": 0.4152, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.21817351076589184, | |
| "eval_loss": 0.24392470717430115, | |
| "eval_runtime": 13.9841, | |
| "eval_samples_per_second": 68.65, | |
| "eval_steps_per_second": 2.145, | |
| "eval_wer": 0.4166666666666667, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 5.08, | |
| "learning_rate": 3.66e-05, | |
| "loss": 0.364, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 5.17, | |
| "learning_rate": 3.72e-05, | |
| "loss": 0.4115, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 5.25, | |
| "learning_rate": 3.7800000000000004e-05, | |
| "loss": 0.3769, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 5.33, | |
| "learning_rate": 3.8400000000000005e-05, | |
| "loss": 0.3635, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 5.42, | |
| "learning_rate": 3.9e-05, | |
| "loss": 0.3743, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 5.5, | |
| "learning_rate": 3.96e-05, | |
| "loss": 0.3188, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 5.58, | |
| "learning_rate": 4.02e-05, | |
| "loss": 0.3608, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 5.67, | |
| "learning_rate": 4.08e-05, | |
| "loss": 0.3285, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 5.75, | |
| "learning_rate": 4.14e-05, | |
| "loss": 0.2964, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 5.83, | |
| "learning_rate": 4.2e-05, | |
| "loss": 0.2799, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 5.92, | |
| "learning_rate": 4.26e-05, | |
| "loss": 0.3272, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 4.32e-05, | |
| "loss": 0.3159, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.17619509966303043, | |
| "eval_loss": 0.13585154712200165, | |
| "eval_runtime": 13.8057, | |
| "eval_samples_per_second": 69.536, | |
| "eval_steps_per_second": 2.173, | |
| "eval_wer": 0.3083708708708709, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 6.08, | |
| "learning_rate": 4.38e-05, | |
| "loss": 0.3256, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 6.17, | |
| "learning_rate": 4.44e-05, | |
| "loss": 0.2651, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "learning_rate": 4.5e-05, | |
| "loss": 0.2502, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 6.33, | |
| "learning_rate": 4.5600000000000004e-05, | |
| "loss": 0.2632, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 6.42, | |
| "learning_rate": 4.6200000000000005e-05, | |
| "loss": 0.2412, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "learning_rate": 4.6800000000000006e-05, | |
| "loss": 0.2871, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 6.58, | |
| "learning_rate": 4.74e-05, | |
| "loss": 0.2409, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 6.67, | |
| "learning_rate": 4.8e-05, | |
| "loss": 0.2091, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 6.75, | |
| "learning_rate": 4.86e-05, | |
| "loss": 0.2677, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 6.83, | |
| "learning_rate": 4.9199999999999997e-05, | |
| "loss": 0.2109, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 6.92, | |
| "learning_rate": 4.98e-05, | |
| "loss": 0.1886, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "learning_rate": 5.04e-05, | |
| "loss": 0.2425, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.15089382603232623, | |
| "eval_loss": 0.07371211796998978, | |
| "eval_runtime": 13.9128, | |
| "eval_samples_per_second": 69.001, | |
| "eval_steps_per_second": 2.156, | |
| "eval_wer": 0.25225225225225223, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 7.08, | |
| "learning_rate": 5.1e-05, | |
| "loss": 0.2069, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 7.17, | |
| "learning_rate": 5.16e-05, | |
| "loss": 0.1888, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 7.25, | |
| "learning_rate": 5.22e-05, | |
| "loss": 0.1926, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 7.33, | |
| "learning_rate": 5.28e-05, | |
| "loss": 0.1794, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 7.42, | |
| "learning_rate": 5.3400000000000004e-05, | |
| "loss": 0.2078, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "learning_rate": 5.4000000000000005e-05, | |
| "loss": 0.1646, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 7.58, | |
| "learning_rate": 5.4600000000000006e-05, | |
| "loss": 0.1417, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 7.67, | |
| "learning_rate": 5.520000000000001e-05, | |
| "loss": 0.1454, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 7.75, | |
| "learning_rate": 5.58e-05, | |
| "loss": 0.173, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 7.83, | |
| "learning_rate": 5.6399999999999995e-05, | |
| "loss": 0.1449, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 7.92, | |
| "learning_rate": 5.6999999999999996e-05, | |
| "loss": 0.1359, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 5.76e-05, | |
| "loss": 0.1921, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.16083157233422812, | |
| "eval_loss": 0.045883145183324814, | |
| "eval_runtime": 13.9631, | |
| "eval_samples_per_second": 68.753, | |
| "eval_steps_per_second": 2.149, | |
| "eval_wer": 0.22128378378378377, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 960, | |
| "total_flos": 2.451994748193635e+18, | |
| "train_loss": 2.0509253946443398, | |
| "train_runtime": 1630.9039, | |
| "train_samples_per_second": 18.836, | |
| "train_steps_per_second": 0.589 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 960, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 2.451994748193635e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |