| { | |
| "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": 9.826288456481726e-05, | |
| "loss": 37.9017, | |
| "step": 4270 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.8080713748931885, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.7466, | |
| "eval_samples_per_second": 27.109, | |
| "eval_steps_per_second": 3.485, | |
| "step": 4270 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 9.124539836970814e-05, | |
| "loss": 3.3383, | |
| "step": 8540 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.6899821758270264, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.6904, | |
| "eval_samples_per_second": 27.307, | |
| "eval_steps_per_second": 3.511, | |
| "step": 8540 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.4227912174599e-05, | |
| "loss": 3.2564, | |
| "step": 12810 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 0.9982683982683983, | |
| "eval_loss": 3.388317823410034, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.7217, | |
| "eval_samples_per_second": 27.196, | |
| "eval_steps_per_second": 3.497, | |
| "step": 12810 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 7.721042597948987e-05, | |
| "loss": 3.104, | |
| "step": 17080 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.9939393939393939, | |
| "eval_loss": 3.202686071395874, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.9831, | |
| "eval_samples_per_second": 26.306, | |
| "eval_steps_per_second": 3.382, | |
| "step": 17080 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 7.019622666316067e-05, | |
| "loss": 2.997, | |
| "step": 21350 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.9904761904761905, | |
| "eval_loss": 3.177474021911621, | |
| "eval_new_wer": 0.9904761904761905, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.7653, | |
| "eval_samples_per_second": 27.043, | |
| "eval_steps_per_second": 3.477, | |
| "step": 21350 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 6.318202734683146e-05, | |
| "loss": 2.9391, | |
| "step": 25620 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.9861471861471861, | |
| "eval_loss": 3.1568691730499268, | |
| "eval_new_wer": 0.9809523809523809, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.7615, | |
| "eval_samples_per_second": 27.057, | |
| "eval_steps_per_second": 3.479, | |
| "step": 25620 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "learning_rate": 5.616782803050224e-05, | |
| "loss": 2.8989, | |
| "step": 29890 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.9766233766233766, | |
| "eval_loss": 3.164285182952881, | |
| "eval_new_wer": 0.9714285714285714, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.8832, | |
| "eval_samples_per_second": 26.639, | |
| "eval_steps_per_second": 3.425, | |
| "step": 29890 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 4.9153628714173024e-05, | |
| "loss": 2.8731, | |
| "step": 34160 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.9670995670995671, | |
| "eval_loss": 3.1436803340911865, | |
| "eval_new_wer": 0.9523809523809523, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.7748, | |
| "eval_samples_per_second": 27.01, | |
| "eval_steps_per_second": 3.473, | |
| "step": 34160 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 4.213778595845386e-05, | |
| "loss": 2.8549, | |
| "step": 38430 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.9645021645021645, | |
| "eval_loss": 3.1454625129699707, | |
| "eval_new_wer": 0.9428571428571428, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.8309, | |
| "eval_samples_per_second": 26.817, | |
| "eval_steps_per_second": 3.448, | |
| "step": 38430 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "learning_rate": 3.512358664212464e-05, | |
| "loss": 2.8407, | |
| "step": 42700 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.9627705627705627, | |
| "eval_loss": 3.1407470703125, | |
| "eval_new_wer": 0.9380952380952381, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.787, | |
| "eval_samples_per_second": 26.968, | |
| "eval_steps_per_second": 3.467, | |
| "step": 42700 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "learning_rate": 2.8107743886405467e-05, | |
| "loss": 2.8301, | |
| "step": 46970 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 0.9601731601731601, | |
| "eval_loss": 3.141080141067505, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.8547, | |
| "eval_samples_per_second": 26.735, | |
| "eval_steps_per_second": 3.437, | |
| "step": 46970 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 2.1093544570076258e-05, | |
| "loss": 2.8215, | |
| "step": 51240 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.9601731601731601, | |
| "eval_loss": 3.1361286640167236, | |
| "eval_new_wer": 0.9380952380952381, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.7971, | |
| "eval_samples_per_second": 26.933, | |
| "eval_steps_per_second": 3.463, | |
| "step": 51240 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "learning_rate": 1.4079345253747043e-05, | |
| "loss": 2.8178, | |
| "step": 55510 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 0.9567099567099567, | |
| "eval_loss": 3.1265242099761963, | |
| "eval_new_wer": 0.9285714285714286, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.8518, | |
| "eval_samples_per_second": 26.746, | |
| "eval_steps_per_second": 3.439, | |
| "step": 55510 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "learning_rate": 7.065145937417829e-06, | |
| "loss": 2.816, | |
| "step": 59780 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 0.9575757575757575, | |
| "eval_loss": 3.1279547214508057, | |
| "eval_new_wer": 0.9238095238095239, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.6197, | |
| "eval_samples_per_second": 24.363, | |
| "eval_steps_per_second": 3.132, | |
| "step": 59780 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "learning_rate": 4.930318169865895e-08, | |
| "loss": 2.8106, | |
| "step": 64050 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 0.9593073593073593, | |
| "eval_loss": 3.131213426589966, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5832, | |
| "eval_samples_per_second": 24.466, | |
| "eval_steps_per_second": 3.146, | |
| "step": 64050 | |
| } | |
| ], | |
| "max_steps": 64050, | |
| "num_train_epochs": 15, | |
| "total_flos": 5.079400634281409e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |