| { | |
| "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": 9.82612411254273e-05, | |
| "loss": 41.6673, | |
| "step": 4270 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.6662509441375732, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.1005, | |
| "eval_samples_per_second": 25.924, | |
| "eval_steps_per_second": 3.333, | |
| "step": 4270 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 9.124375493031818e-05, | |
| "loss": 3.3366, | |
| "step": 8540 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.467925548553467, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.5159, | |
| "eval_samples_per_second": 24.66, | |
| "eval_steps_per_second": 3.171, | |
| "step": 8540 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.422626873520904e-05, | |
| "loss": 3.2357, | |
| "step": 12810 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 1.0, | |
| "eval_loss": 3.2729125022888184, | |
| "eval_new_wer": 1.0, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 7.8374, | |
| "eval_samples_per_second": 26.794, | |
| "eval_steps_per_second": 3.445, | |
| "step": 12810 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 7.721042597948987e-05, | |
| "loss": 3.0695, | |
| "step": 17080 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 0.9965367965367965, | |
| "eval_loss": 3.2063045501708984, | |
| "eval_new_wer": 0.9952380952380953, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.0285, | |
| "eval_samples_per_second": 26.157, | |
| "eval_steps_per_second": 3.363, | |
| "step": 17080 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 7.019622666316067e-05, | |
| "loss": 2.9729, | |
| "step": 21350 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.987012987012987, | |
| "eval_loss": 3.182396411895752, | |
| "eval_new_wer": 0.9809523809523809, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.1314, | |
| "eval_samples_per_second": 25.826, | |
| "eval_steps_per_second": 3.32, | |
| "step": 21350 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 6.318202734683146e-05, | |
| "loss": 2.9161, | |
| "step": 25620 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 0.9766233766233766, | |
| "eval_loss": 3.1507506370544434, | |
| "eval_new_wer": 0.9619047619047619, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.2738, | |
| "eval_samples_per_second": 25.381, | |
| "eval_steps_per_second": 3.263, | |
| "step": 25620 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "learning_rate": 5.616782803050224e-05, | |
| "loss": 2.8792, | |
| "step": 29890 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 0.9688311688311688, | |
| "eval_loss": 3.118381977081299, | |
| "eval_new_wer": 0.9476190476190476, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.1799, | |
| "eval_samples_per_second": 25.673, | |
| "eval_steps_per_second": 3.301, | |
| "step": 29890 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 4.9153628714173024e-05, | |
| "loss": 2.8532, | |
| "step": 34160 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 0.9679653679653679, | |
| "eval_loss": 3.119230031967163, | |
| "eval_new_wer": 0.9476190476190476, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.122, | |
| "eval_samples_per_second": 25.856, | |
| "eval_steps_per_second": 3.324, | |
| "step": 34160 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 4.213942939784381e-05, | |
| "loss": 2.8319, | |
| "step": 38430 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 0.9679653679653679, | |
| "eval_loss": 3.1226253509521484, | |
| "eval_new_wer": 0.9476190476190476, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.203, | |
| "eval_samples_per_second": 25.6, | |
| "eval_steps_per_second": 3.291, | |
| "step": 38430 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "learning_rate": 3.512358664212464e-05, | |
| "loss": 2.8189, | |
| "step": 42700 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 0.9636363636363636, | |
| "eval_loss": 3.1021082401275635, | |
| "eval_new_wer": 0.9428571428571428, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.2497, | |
| "eval_samples_per_second": 25.455, | |
| "eval_steps_per_second": 3.273, | |
| "step": 42700 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "learning_rate": 2.8107743886405467e-05, | |
| "loss": 2.8081, | |
| "step": 46970 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 0.9584415584415584, | |
| "eval_loss": 3.1021568775177, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.1922, | |
| "eval_samples_per_second": 25.634, | |
| "eval_steps_per_second": 3.296, | |
| "step": 46970 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "learning_rate": 2.1095188009466215e-05, | |
| "loss": 2.8027, | |
| "step": 51240 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 0.9575757575757575, | |
| "eval_loss": 3.1064510345458984, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.1815, | |
| "eval_samples_per_second": 25.668, | |
| "eval_steps_per_second": 3.3, | |
| "step": 51240 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "learning_rate": 1.4080988693136998e-05, | |
| "loss": 2.7939, | |
| "step": 55510 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 0.9575757575757575, | |
| "eval_loss": 3.102414608001709, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 9.239, | |
| "eval_samples_per_second": 22.73, | |
| "eval_steps_per_second": 2.922, | |
| "step": 55510 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "learning_rate": 7.066789376807783e-06, | |
| "loss": 2.7917, | |
| "step": 59780 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 0.9584415584415584, | |
| "eval_loss": 3.100308895111084, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 9.2123, | |
| "eval_samples_per_second": 22.796, | |
| "eval_steps_per_second": 2.931, | |
| "step": 59780 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "learning_rate": 5.094662108861425e-08, | |
| "loss": 2.7903, | |
| "step": 64050 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 0.9575757575757575, | |
| "eval_loss": 3.104132890701294, | |
| "eval_new_wer": 0.9333333333333333, | |
| "eval_old_wer": 1.0, | |
| "eval_runtime": 8.8255, | |
| "eval_samples_per_second": 23.795, | |
| "eval_steps_per_second": 3.059, | |
| "step": 64050 | |
| } | |
| ], | |
| "max_steps": 64050, | |
| "num_train_epochs": 15, | |
| "total_flos": 5.076674134512955e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |