| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "global_step": 64050, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 9.825795424664739e-05, | |
| "loss": 7.898, | |
| "step": 4270 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.18354978354978355, | |
| "eval_loss": 0.892418622970581, | |
| "eval_new_wer": 0.1, | |
| "eval_old_wer": 0.4714285714285714, | |
| "eval_runtime": 8.5895, | |
| "eval_samples_per_second": 24.448, | |
| "eval_steps_per_second": 3.143, | |
| "step": 4270 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 9.124046805153827e-05, | |
| "loss": 0.9517, | |
| "step": 8540 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.14545454545454545, | |
| "eval_loss": 0.8618147373199463, | |
| "eval_new_wer": 0.12857142857142856, | |
| "eval_old_wer": 0.30952380952380953, | |
| "eval_runtime": 8.4071, | |
| "eval_samples_per_second": 24.979, | |
| "eval_steps_per_second": 3.212, | |
| "step": 8540 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.422626873520904e-05, | |
| "loss": 0.6119, | |
| "step": 12810 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.15064935064935064, | |
| "eval_loss": 0.8806686997413635, | |
| "eval_new_wer": 0.12380952380952381, | |
| "eval_old_wer": 0.2904761904761905, | |
| "eval_runtime": 8.4758, | |
| "eval_samples_per_second": 24.776, | |
| "eval_steps_per_second": 3.186, | |
| "step": 12810 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 7.721206941887984e-05, | |
| "loss": 0.4562, | |
| "step": 17080 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.0987012987012987, | |
| "eval_loss": 0.6651872992515564, | |
| "eval_new_wer": 0.09047619047619047, | |
| "eval_old_wer": 0.2, | |
| "eval_runtime": 8.491, | |
| "eval_samples_per_second": 24.732, | |
| "eval_steps_per_second": 3.18, | |
| "step": 17080 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 7.019787010255062e-05, | |
| "loss": 0.3661, | |
| "step": 21350 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.13766233766233765, | |
| "eval_loss": 1.2836520671844482, | |
| "eval_new_wer": 0.19047619047619047, | |
| "eval_old_wer": 0.2619047619047619, | |
| "eval_runtime": 8.4952, | |
| "eval_samples_per_second": 24.72, | |
| "eval_steps_per_second": 3.178, | |
| "step": 21350 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 6.318202734683146e-05, | |
| "loss": 0.2908, | |
| "step": 25620 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.10735930735930736, | |
| "eval_loss": 0.6425668597221375, | |
| "eval_new_wer": 0.10952380952380952, | |
| "eval_old_wer": 0.18571428571428572, | |
| "eval_runtime": 8.4544, | |
| "eval_samples_per_second": 24.839, | |
| "eval_steps_per_second": 3.194, | |
| "step": 25620 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "learning_rate": 5.616782803050224e-05, | |
| "loss": 0.2357, | |
| "step": 29890 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.15411255411255412, | |
| "eval_loss": 1.2289475202560425, | |
| "eval_new_wer": 0.1619047619047619, | |
| "eval_old_wer": 0.24761904761904763, | |
| "eval_runtime": 8.4851, | |
| "eval_samples_per_second": 24.749, | |
| "eval_steps_per_second": 3.182, | |
| "step": 29890 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 4.915198527478307e-05, | |
| "loss": 0.1926, | |
| "step": 34160 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.12727272727272726, | |
| "eval_loss": 1.098648190498352, | |
| "eval_new_wer": 0.14761904761904762, | |
| "eval_old_wer": 0.23809523809523808, | |
| "eval_runtime": 8.5184, | |
| "eval_samples_per_second": 24.653, | |
| "eval_steps_per_second": 3.17, | |
| "step": 34160 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 4.213778595845386e-05, | |
| "loss": 0.1601, | |
| "step": 38430 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.14805194805194805, | |
| "eval_loss": 1.2215726375579834, | |
| "eval_new_wer": 0.18095238095238095, | |
| "eval_old_wer": 0.24285714285714285, | |
| "eval_runtime": 8.367, | |
| "eval_samples_per_second": 25.098, | |
| "eval_steps_per_second": 3.227, | |
| "step": 38430 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "learning_rate": 3.5121943202734684e-05, | |
| "loss": 0.1317, | |
| "step": 42700 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.12034632034632034, | |
| "eval_loss": 0.9588236212730408, | |
| "eval_new_wer": 0.14761904761904762, | |
| "eval_old_wer": 0.20476190476190476, | |
| "eval_runtime": 8.5613, | |
| "eval_samples_per_second": 24.529, | |
| "eval_steps_per_second": 3.154, | |
| "step": 42700 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "learning_rate": 2.8106100447015514e-05, | |
| "loss": 0.1088, | |
| "step": 46970 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 0.09610389610389611, | |
| "eval_loss": 0.7578983902931213, | |
| "eval_new_wer": 0.10476190476190476, | |
| "eval_old_wer": 0.1761904761904762, | |
| "eval_runtime": 8.5378, | |
| "eval_samples_per_second": 24.597, | |
| "eval_steps_per_second": 3.162, | |
| "step": 46970 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 2.1093544570076258e-05, | |
| "loss": 0.0892, | |
| "step": 51240 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.10043290043290043, | |
| "eval_loss": 0.7190653681755066, | |
| "eval_new_wer": 0.11428571428571428, | |
| "eval_old_wer": 0.16666666666666666, | |
| "eval_runtime": 8.7441, | |
| "eval_samples_per_second": 24.016, | |
| "eval_steps_per_second": 3.088, | |
| "step": 51240 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "learning_rate": 1.4079345253747043e-05, | |
| "loss": 0.0714, | |
| "step": 55510 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 0.08398268398268398, | |
| "eval_loss": 0.4884723424911499, | |
| "eval_new_wer": 0.09523809523809523, | |
| "eval_old_wer": 0.14285714285714285, | |
| "eval_runtime": 8.4874, | |
| "eval_samples_per_second": 24.743, | |
| "eval_steps_per_second": 3.181, | |
| "step": 55510 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "learning_rate": 7.063502498027873e-06, | |
| "loss": 0.0573, | |
| "step": 59780 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 0.09177489177489177, | |
| "eval_loss": 0.5790263414382935, | |
| "eval_new_wer": 0.10952380952380952, | |
| "eval_old_wer": 0.14285714285714285, | |
| "eval_runtime": 8.6001, | |
| "eval_samples_per_second": 24.418, | |
| "eval_steps_per_second": 3.139, | |
| "step": 59780 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "learning_rate": 4.7659742308703654e-08, | |
| "loss": 0.0469, | |
| "step": 64050 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 0.09004329004329005, | |
| "eval_loss": 0.6429840922355652, | |
| "eval_new_wer": 0.09523809523809523, | |
| "eval_old_wer": 0.14285714285714285, | |
| "eval_runtime": 8.622, | |
| "eval_samples_per_second": 24.356, | |
| "eval_steps_per_second": 3.132, | |
| "step": 64050 | |
| } | |
| ], | |
| "max_steps": 64050, | |
| "num_train_epochs": 15, | |
| "total_flos": 4.852750717448731e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |