| { | |
| "best_metric": 67.83942963422194, | |
| "best_model_checkpoint": "./checkpoint-300", | |
| "epoch": 42.857142857142854, | |
| "global_step": 300, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 2.875862068965517e-07, | |
| "loss": 2.755, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 7.14, | |
| "learning_rate": 2.6172413793103444e-07, | |
| "loss": 2.0968, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 10.71, | |
| "learning_rate": 2.3586206896551725e-07, | |
| "loss": 1.703, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 14.29, | |
| "learning_rate": 2.0999999999999997e-07, | |
| "loss": 1.4728, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 14.29, | |
| "eval_loss": 1.4428775310516357, | |
| "eval_runtime": 1406.9013, | |
| "eval_samples_per_second": 0.364, | |
| "eval_steps_per_second": 0.023, | |
| "eval_wer": 83.59423434593924, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 17.86, | |
| "learning_rate": 1.8413793103448276e-07, | |
| "loss": 1.2819, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 21.43, | |
| "learning_rate": 1.5827586206896554e-07, | |
| "loss": 1.1744, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "learning_rate": 1.3241379310344827e-07, | |
| "loss": 1.0885, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 28.57, | |
| "learning_rate": 1.0655172413793102e-07, | |
| "loss": 1.0344, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 28.57, | |
| "eval_loss": 1.158418893814087, | |
| "eval_runtime": 1566.2048, | |
| "eval_samples_per_second": 0.327, | |
| "eval_steps_per_second": 0.02, | |
| "eval_wer": 69.97055176689399, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 32.14, | |
| "learning_rate": 8.068965517241379e-08, | |
| "loss": 0.9889, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 35.71, | |
| "learning_rate": 5.482758620689655e-08, | |
| "loss": 0.9647, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 39.29, | |
| "learning_rate": 2.8965517241379308e-08, | |
| "loss": 0.9435, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 42.86, | |
| "learning_rate": 3.103448275862069e-09, | |
| "loss": 0.9318, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 42.86, | |
| "eval_loss": 1.1061279773712158, | |
| "eval_runtime": 1528.8642, | |
| "eval_samples_per_second": 0.335, | |
| "eval_steps_per_second": 0.021, | |
| "eval_wer": 67.83942963422194, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 42.86, | |
| "step": 300, | |
| "total_flos": 1.886690232041472e+19, | |
| "train_loss": 0.31906700134277344, | |
| "train_runtime": 2410.2404, | |
| "train_samples_per_second": 7.966, | |
| "train_steps_per_second": 0.124 | |
| } | |
| ], | |
| "max_steps": 300, | |
| "num_train_epochs": 43, | |
| "total_flos": 1.886690232041472e+19, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |