| { | |
| "best_metric": 0.5780478119850159, | |
| "best_model_checkpoint": "whisper-reg-ben/checkpoint-5394", | |
| "epoch": 2.0, | |
| "eval_steps": 500, | |
| "global_step": 5394, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 3.480727434158325, | |
| "learning_rate": 4.623158292406498e-05, | |
| "loss": 0.2145, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 5.088292121887207, | |
| "learning_rate": 4.151870041556479e-05, | |
| "loss": 0.2333, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 3.8999531269073486, | |
| "learning_rate": 3.679637325273895e-05, | |
| "loss": 0.2302, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 3.2788655757904053, | |
| "learning_rate": 3.2074046089913114e-05, | |
| "loss": 0.2293, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 2.2892401218414307, | |
| "learning_rate": 2.735171892708727e-05, | |
| "loss": 0.2166, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 0.3217996656894684, | |
| "eval_loss": 0.2115698903799057, | |
| "eval_runtime": 9048.1139, | |
| "eval_samples_per_second": 0.298, | |
| "eval_steps_per_second": 0.075, | |
| "eval_wer": 0.6381270885467529, | |
| "step": 2697 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "grad_norm": 3.279925584793091, | |
| "learning_rate": 2.262939176426143e-05, | |
| "loss": 0.1708, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 2.526437997817993, | |
| "learning_rate": 1.790706460143559e-05, | |
| "loss": 0.1369, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.48, | |
| "grad_norm": 3.6526429653167725, | |
| "learning_rate": 1.3184737438609747e-05, | |
| "loss": 0.1288, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 2.9243693351745605, | |
| "learning_rate": 8.462410275783907e-06, | |
| "loss": 0.1239, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.85, | |
| "grad_norm": 3.0473432540893555, | |
| "learning_rate": 3.749527767283718e-06, | |
| "loss": 0.1149, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 0.28478333353996277, | |
| "eval_loss": 0.18326056003570557, | |
| "eval_runtime": 9056.1912, | |
| "eval_samples_per_second": 0.298, | |
| "eval_steps_per_second": 0.075, | |
| "eval_wer": 0.5780478119850159, | |
| "step": 5394 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5394, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 500, | |
| "total_flos": 6.2247870849024e+18, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |