| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "global_step": 4830, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 9.83874482458052e-05, | |
| "loss": 22.5987, | |
| "step": 322 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.259235143661499, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.6332, | |
| "eval_samples_per_second": 24.325, | |
| "eval_steps_per_second": 3.127, | |
| "step": 322 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 9.13706689910656e-05, | |
| "loss": 3.2658, | |
| "step": 644 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.177554130554199, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5905, | |
| "eval_samples_per_second": 24.446, | |
| "eval_steps_per_second": 3.143, | |
| "step": 644 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.4353889736326e-05, | |
| "loss": 3.0366, | |
| "step": 966 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.093811511993408, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 9.0154, | |
| "eval_samples_per_second": 23.293, | |
| "eval_steps_per_second": 2.995, | |
| "step": 966 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 7.73371104815864e-05, | |
| "loss": 2.635, | |
| "step": 1288 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.0094852447509766, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5468, | |
| "eval_samples_per_second": 24.571, | |
| "eval_steps_per_second": 3.159, | |
| "step": 1288 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 7.032033122684682e-05, | |
| "loss": 2.0662, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.9774891774891775, | |
| "eval_loss": 2.8005685806274414, | |
| "eval_new_wer": 0.9952380952380953, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.9981, | |
| "eval_samples_per_second": 23.338, | |
| "eval_steps_per_second": 3.001, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 6.330355197210721e-05, | |
| "loss": 1.5835, | |
| "step": 1932 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.7350649350649351, | |
| "eval_loss": 2.7513883113861084, | |
| "eval_new_wer": 0.5238095238095238, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5668, | |
| "eval_samples_per_second": 24.513, | |
| "eval_steps_per_second": 3.152, | |
| "step": 1932 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "learning_rate": 5.628677271736762e-05, | |
| "loss": 1.204, | |
| "step": 2254 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.7177489177489178, | |
| "eval_loss": 3.089792013168335, | |
| "eval_new_wer": 0.5666666666666667, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.7621, | |
| "eval_samples_per_second": 23.967, | |
| "eval_steps_per_second": 3.081, | |
| "step": 2254 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 4.9269993462628025e-05, | |
| "loss": 0.8525, | |
| "step": 2576 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.625974025974026, | |
| "eval_loss": 3.1094284057617188, | |
| "eval_new_wer": 0.35714285714285715, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5847, | |
| "eval_samples_per_second": 24.462, | |
| "eval_steps_per_second": 3.145, | |
| "step": 2576 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 4.225321420788843e-05, | |
| "loss": 0.6166, | |
| "step": 2898 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.6337662337662338, | |
| "eval_loss": 3.5997140407562256, | |
| "eval_new_wer": 0.35714285714285715, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5651, | |
| "eval_samples_per_second": 24.518, | |
| "eval_steps_per_second": 3.152, | |
| "step": 2898 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "learning_rate": 3.5236434953148835e-05, | |
| "loss": 0.4461, | |
| "step": 3220 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.6562770562770562, | |
| "eval_loss": 3.8030993938446045, | |
| "eval_new_wer": 0.3523809523809524, | |
| "eval_old_wer": 0.9952380952380953, | |
| "eval_runtime": 8.652, | |
| "eval_samples_per_second": 24.272, | |
| "eval_steps_per_second": 3.121, | |
| "step": 3220 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "learning_rate": 2.8219655698409243e-05, | |
| "loss": 0.3257, | |
| "step": 3542 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 0.6060606060606061, | |
| "eval_loss": 4.010402202606201, | |
| "eval_new_wer": 0.3238095238095238, | |
| "eval_old_wer": 0.9857142857142858, | |
| "eval_runtime": 9.2292, | |
| "eval_samples_per_second": 22.754, | |
| "eval_steps_per_second": 2.925, | |
| "step": 3542 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 2.1202876443669645e-05, | |
| "loss": 0.254, | |
| "step": 3864 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.6164502164502165, | |
| "eval_loss": 3.9872653484344482, | |
| "eval_new_wer": 0.3238095238095238, | |
| "eval_old_wer": 0.9857142857142858, | |
| "eval_runtime": 8.5518, | |
| "eval_samples_per_second": 24.556, | |
| "eval_steps_per_second": 3.157, | |
| "step": 3864 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "learning_rate": 1.418609718893005e-05, | |
| "loss": 0.1988, | |
| "step": 4186 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 0.5922077922077922, | |
| "eval_loss": 3.9551496505737305, | |
| "eval_new_wer": 0.2619047619047619, | |
| "eval_old_wer": 0.9714285714285714, | |
| "eval_runtime": 8.0772, | |
| "eval_samples_per_second": 25.999, | |
| "eval_steps_per_second": 3.343, | |
| "step": 4186 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "learning_rate": 7.1693179341904555e-06, | |
| "loss": 0.1663, | |
| "step": 4508 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 0.6095238095238096, | |
| "eval_loss": 4.294742107391357, | |
| "eval_new_wer": 0.2904761904761905, | |
| "eval_old_wer": 0.9809523809523809, | |
| "eval_runtime": 8.1055, | |
| "eval_samples_per_second": 25.908, | |
| "eval_steps_per_second": 3.331, | |
| "step": 4508 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "learning_rate": 1.5253867945086075e-07, | |
| "loss": 0.1398, | |
| "step": 4830 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 0.6207792207792208, | |
| "eval_loss": 4.198614120483398, | |
| "eval_new_wer": 0.319047619047619, | |
| "eval_old_wer": 0.9904761904761905, | |
| "eval_runtime": 8.1622, | |
| "eval_samples_per_second": 25.728, | |
| "eval_steps_per_second": 3.308, | |
| "step": 4830 | |
| } | |
| ], | |
| "max_steps": 4830, | |
| "num_train_epochs": 15, | |
| "total_flos": 3.586056332009567e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |