| { | |
| "best_metric": 8.60641891891892, | |
| "best_model_checkpoint": "./checkpoint-9000", | |
| "epoch": 10.97694840834248, | |
| "global_step": 10000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 1.94e-06, | |
| "loss": 0.8124, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 3.94e-06, | |
| "loss": 0.3235, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 5.94e-06, | |
| "loss": 0.2396, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 7.94e-06, | |
| "loss": 0.2138, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 9.940000000000001e-06, | |
| "loss": 0.1972, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 9.897894736842107e-06, | |
| "loss": 0.1885, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 9.79263157894737e-06, | |
| "loss": 0.1792, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 9.687368421052632e-06, | |
| "loss": 0.174, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 9.582105263157897e-06, | |
| "loss": 0.1669, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "learning_rate": 9.476842105263158e-06, | |
| "loss": 0.1106, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_cer": 6.4784271464430745, | |
| "eval_loss": 0.18272116780281067, | |
| "eval_runtime": 4332.708, | |
| "eval_samples_per_second": 1.063, | |
| "eval_steps_per_second": 0.066, | |
| "eval_wer": 10.347972972972972, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 9.371578947368421e-06, | |
| "loss": 0.1083, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.32, | |
| "learning_rate": 9.266315789473685e-06, | |
| "loss": 0.1043, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 9.161052631578948e-06, | |
| "loss": 0.1016, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.54, | |
| "learning_rate": 9.055789473684211e-06, | |
| "loss": 0.1023, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.65, | |
| "learning_rate": 8.950526315789474e-06, | |
| "loss": 0.101, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.76, | |
| "learning_rate": 8.845263157894738e-06, | |
| "loss": 0.099, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 8.740000000000001e-06, | |
| "loss": 0.1005, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 8.634736842105264e-06, | |
| "loss": 0.1002, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.09, | |
| "learning_rate": 8.529473684210528e-06, | |
| "loss": 0.0601, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 8.42421052631579e-06, | |
| "loss": 0.0487, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_cer": 5.912657213073819, | |
| "eval_loss": 0.17993855476379395, | |
| "eval_runtime": 4312.9082, | |
| "eval_samples_per_second": 1.067, | |
| "eval_steps_per_second": 0.067, | |
| "eval_wer": 9.47635135135135, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.31, | |
| "learning_rate": 8.318947368421052e-06, | |
| "loss": 0.0508, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.41, | |
| "learning_rate": 8.213684210526316e-06, | |
| "loss": 0.0501, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.52, | |
| "learning_rate": 8.10842105263158e-06, | |
| "loss": 0.0506, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.63, | |
| "learning_rate": 8.003157894736842e-06, | |
| "loss": 0.0498, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.74, | |
| "learning_rate": 7.897894736842106e-06, | |
| "loss": 0.0486, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.85, | |
| "learning_rate": 7.792631578947369e-06, | |
| "loss": 0.0513, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.96, | |
| "learning_rate": 7.687368421052632e-06, | |
| "loss": 0.0504, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "learning_rate": 7.582105263157895e-06, | |
| "loss": 0.0307, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.18, | |
| "learning_rate": 7.476842105263159e-06, | |
| "loss": 0.0239, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 7.371578947368422e-06, | |
| "loss": 0.0243, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "eval_cer": 5.806862835289325, | |
| "eval_loss": 0.19502821564674377, | |
| "eval_runtime": 4299.7561, | |
| "eval_samples_per_second": 1.071, | |
| "eval_steps_per_second": 0.067, | |
| "eval_wer": 9.21114864864865, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 7.266315789473684e-06, | |
| "loss": 0.024, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 3.51, | |
| "learning_rate": 7.161052631578948e-06, | |
| "loss": 0.0252, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 3.62, | |
| "learning_rate": 7.055789473684212e-06, | |
| "loss": 0.0244, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 3.73, | |
| "learning_rate": 6.950526315789474e-06, | |
| "loss": 0.0249, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 3.84, | |
| "learning_rate": 6.845263157894737e-06, | |
| "loss": 0.0247, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 3.95, | |
| "learning_rate": 6.740000000000001e-06, | |
| "loss": 0.0245, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 4.06, | |
| "learning_rate": 6.634736842105264e-06, | |
| "loss": 0.0173, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 4.17, | |
| "learning_rate": 6.529473684210526e-06, | |
| "loss": 0.0106, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 4.28, | |
| "learning_rate": 6.42421052631579e-06, | |
| "loss": 0.0116, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 4.39, | |
| "learning_rate": 6.318947368421054e-06, | |
| "loss": 0.0106, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.39, | |
| "eval_cer": 5.575560841623846, | |
| "eval_loss": 0.21126572787761688, | |
| "eval_runtime": 4218.4134, | |
| "eval_samples_per_second": 1.091, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.971283783783784, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 6.213684210526316e-06, | |
| "loss": 0.0115, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 4.61, | |
| "learning_rate": 6.108421052631579e-06, | |
| "loss": 0.0105, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "learning_rate": 6.003157894736843e-06, | |
| "loss": 0.0104, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 4.83, | |
| "learning_rate": 5.897894736842106e-06, | |
| "loss": 0.0123, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 4.94, | |
| "learning_rate": 5.7926315789473685e-06, | |
| "loss": 0.0116, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 5.05, | |
| "learning_rate": 5.687368421052633e-06, | |
| "loss": 0.0079, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 5.16, | |
| "learning_rate": 5.582105263157895e-06, | |
| "loss": 0.0053, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 5.27, | |
| "learning_rate": 5.476842105263158e-06, | |
| "loss": 0.0052, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 5.38, | |
| "learning_rate": 5.371578947368421e-06, | |
| "loss": 0.0053, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 5.49, | |
| "learning_rate": 5.266315789473685e-06, | |
| "loss": 0.0054, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.49, | |
| "eval_cer": 5.404055670184384, | |
| "eval_loss": 0.23248465359210968, | |
| "eval_runtime": 4237.1053, | |
| "eval_samples_per_second": 1.087, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.64695945945946, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.6, | |
| "learning_rate": 5.161052631578948e-06, | |
| "loss": 0.0053, | |
| "step": 5100 | |
| }, | |
| { | |
| "epoch": 5.71, | |
| "learning_rate": 5.0557894736842105e-06, | |
| "loss": 0.0056, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "learning_rate": 4.950526315789474e-06, | |
| "loss": 0.005, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 5.93, | |
| "learning_rate": 4.845263157894737e-06, | |
| "loss": 0.0054, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 6.04, | |
| "learning_rate": 4.74e-06, | |
| "loss": 0.0045, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "learning_rate": 4.634736842105264e-06, | |
| "loss": 0.0029, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 6.26, | |
| "learning_rate": 4.529473684210527e-06, | |
| "loss": 0.0024, | |
| "step": 5700 | |
| }, | |
| { | |
| "epoch": 6.37, | |
| "learning_rate": 4.424210526315789e-06, | |
| "loss": 0.0025, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 6.48, | |
| "learning_rate": 4.3189473684210535e-06, | |
| "loss": 0.0026, | |
| "step": 5900 | |
| }, | |
| { | |
| "epoch": 6.59, | |
| "learning_rate": 4.213684210526316e-06, | |
| "loss": 0.0031, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.59, | |
| "eval_cer": 5.440853714631165, | |
| "eval_loss": 0.24615277349948883, | |
| "eval_runtime": 4218.0606, | |
| "eval_samples_per_second": 1.091, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.70777027027027, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.7, | |
| "learning_rate": 4.108421052631579e-06, | |
| "loss": 0.0028, | |
| "step": 6100 | |
| }, | |
| { | |
| "epoch": 6.81, | |
| "learning_rate": 4.0031578947368424e-06, | |
| "loss": 0.0026, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 6.92, | |
| "learning_rate": 3.897894736842106e-06, | |
| "loss": 0.0027, | |
| "step": 6300 | |
| }, | |
| { | |
| "epoch": 7.03, | |
| "learning_rate": 3.792631578947369e-06, | |
| "loss": 0.0021, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 7.14, | |
| "learning_rate": 3.687368421052632e-06, | |
| "loss": 0.0014, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 7.24, | |
| "learning_rate": 3.582105263157895e-06, | |
| "loss": 0.0015, | |
| "step": 6600 | |
| }, | |
| { | |
| "epoch": 7.35, | |
| "learning_rate": 3.476842105263158e-06, | |
| "loss": 0.001, | |
| "step": 6700 | |
| }, | |
| { | |
| "epoch": 7.46, | |
| "learning_rate": 3.3715789473684212e-06, | |
| "loss": 0.0015, | |
| "step": 6800 | |
| }, | |
| { | |
| "epoch": 7.57, | |
| "learning_rate": 3.2663157894736845e-06, | |
| "loss": 0.0013, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 7.68, | |
| "learning_rate": 3.1610526315789474e-06, | |
| "loss": 0.0014, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.68, | |
| "eval_cer": 5.484879946379992, | |
| "eval_loss": 0.26076802611351013, | |
| "eval_runtime": 4205.2214, | |
| "eval_samples_per_second": 1.095, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.714527027027028, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 3.055789473684211e-06, | |
| "loss": 0.0013, | |
| "step": 7100 | |
| }, | |
| { | |
| "epoch": 7.9, | |
| "learning_rate": 2.9505263157894735e-06, | |
| "loss": 0.0013, | |
| "step": 7200 | |
| }, | |
| { | |
| "epoch": 8.01, | |
| "learning_rate": 2.845263157894737e-06, | |
| "loss": 0.0018, | |
| "step": 7300 | |
| }, | |
| { | |
| "epoch": 8.12, | |
| "learning_rate": 2.7400000000000004e-06, | |
| "loss": 0.0011, | |
| "step": 7400 | |
| }, | |
| { | |
| "epoch": 8.23, | |
| "learning_rate": 2.6347368421052633e-06, | |
| "loss": 0.0009, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 8.34, | |
| "learning_rate": 2.5294736842105266e-06, | |
| "loss": 0.0008, | |
| "step": 7600 | |
| }, | |
| { | |
| "epoch": 8.45, | |
| "learning_rate": 2.42421052631579e-06, | |
| "loss": 0.0008, | |
| "step": 7700 | |
| }, | |
| { | |
| "epoch": 8.56, | |
| "learning_rate": 2.3189473684210527e-06, | |
| "loss": 0.0008, | |
| "step": 7800 | |
| }, | |
| { | |
| "epoch": 8.67, | |
| "learning_rate": 2.213684210526316e-06, | |
| "loss": 0.0008, | |
| "step": 7900 | |
| }, | |
| { | |
| "epoch": 8.78, | |
| "learning_rate": 2.1084210526315792e-06, | |
| "loss": 0.0009, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.78, | |
| "eval_cer": 5.387627971770643, | |
| "eval_loss": 0.269546777009964, | |
| "eval_runtime": 4205.0269, | |
| "eval_samples_per_second": 1.095, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.630067567567568, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.89, | |
| "learning_rate": 2.003157894736842e-06, | |
| "loss": 0.0007, | |
| "step": 8100 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 1.8978947368421056e-06, | |
| "loss": 0.0009, | |
| "step": 8200 | |
| }, | |
| { | |
| "epoch": 9.11, | |
| "learning_rate": 1.7926315789473686e-06, | |
| "loss": 0.0007, | |
| "step": 8300 | |
| }, | |
| { | |
| "epoch": 9.22, | |
| "learning_rate": 1.6873684210526317e-06, | |
| "loss": 0.0006, | |
| "step": 8400 | |
| }, | |
| { | |
| "epoch": 9.33, | |
| "learning_rate": 1.5821052631578948e-06, | |
| "loss": 0.0005, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 9.44, | |
| "learning_rate": 1.4768421052631578e-06, | |
| "loss": 0.0005, | |
| "step": 8600 | |
| }, | |
| { | |
| "epoch": 9.55, | |
| "learning_rate": 1.3715789473684213e-06, | |
| "loss": 0.0006, | |
| "step": 8700 | |
| }, | |
| { | |
| "epoch": 9.66, | |
| "learning_rate": 1.2663157894736844e-06, | |
| "loss": 0.0005, | |
| "step": 8800 | |
| }, | |
| { | |
| "epoch": 9.77, | |
| "learning_rate": 1.1610526315789474e-06, | |
| "loss": 0.0005, | |
| "step": 8900 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "learning_rate": 1.0557894736842105e-06, | |
| "loss": 0.0004, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "eval_cer": 5.352801251133512, | |
| "eval_loss": 0.2793508470058441, | |
| "eval_runtime": 4203.0281, | |
| "eval_samples_per_second": 1.095, | |
| "eval_steps_per_second": 0.069, | |
| "eval_wer": 8.60641891891892, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.99, | |
| "learning_rate": 9.505263157894738e-07, | |
| "loss": 0.0005, | |
| "step": 9100 | |
| }, | |
| { | |
| "epoch": 10.1, | |
| "learning_rate": 8.452631578947369e-07, | |
| "loss": 0.0005, | |
| "step": 9200 | |
| }, | |
| { | |
| "epoch": 10.21, | |
| "learning_rate": 7.4e-07, | |
| "loss": 0.0003, | |
| "step": 9300 | |
| }, | |
| { | |
| "epoch": 10.32, | |
| "learning_rate": 6.347368421052633e-07, | |
| "loss": 0.0004, | |
| "step": 9400 | |
| }, | |
| { | |
| "epoch": 10.43, | |
| "learning_rate": 5.294736842105263e-07, | |
| "loss": 0.0003, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 10.54, | |
| "learning_rate": 4.242105263157895e-07, | |
| "loss": 0.0004, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 10.65, | |
| "learning_rate": 3.1894736842105263e-07, | |
| "loss": 0.0003, | |
| "step": 9700 | |
| }, | |
| { | |
| "epoch": 10.76, | |
| "learning_rate": 2.136842105263158e-07, | |
| "loss": 0.0003, | |
| "step": 9800 | |
| }, | |
| { | |
| "epoch": 10.87, | |
| "learning_rate": 1.0842105263157895e-07, | |
| "loss": 0.0004, | |
| "step": 9900 | |
| }, | |
| { | |
| "epoch": 10.98, | |
| "learning_rate": 3.1578947368421054e-09, | |
| "loss": 0.0003, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.98, | |
| "eval_cer": 5.469766463839351, | |
| "eval_loss": 0.2836119830608368, | |
| "eval_runtime": 4263.2576, | |
| "eval_samples_per_second": 1.08, | |
| "eval_steps_per_second": 0.068, | |
| "eval_wer": 8.721283783783784, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.98, | |
| "step": 10000, | |
| "total_flos": 3.265323341119488e+20, | |
| "train_loss": 0.043234722255170346, | |
| "train_runtime": 85032.8136, | |
| "train_samples_per_second": 3.763, | |
| "train_steps_per_second": 0.118 | |
| } | |
| ], | |
| "max_steps": 10000, | |
| "num_train_epochs": 11, | |
| "total_flos": 3.265323341119488e+20, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |