|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 48.87983706720978, |
|
"eval_steps": 500, |
|
"global_step": 6000, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 4.07, |
|
"grad_norm": 3.8863396644592285, |
|
"learning_rate": 3.936e-05, |
|
"loss": 3.9012, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 4.07, |
|
"eval_loss": 0.4798698425292969, |
|
"eval_runtime": 164.8654, |
|
"eval_samples_per_second": 19.234, |
|
"eval_steps_per_second": 2.408, |
|
"eval_wer": 0.46691878014085975, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 8.15, |
|
"grad_norm": 1.7655788660049438, |
|
"learning_rate": 7.936000000000001e-05, |
|
"loss": 0.7087, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 8.15, |
|
"eval_loss": 0.331632524728775, |
|
"eval_runtime": 165.1487, |
|
"eval_samples_per_second": 19.201, |
|
"eval_steps_per_second": 2.404, |
|
"eval_wer": 0.3308121985914027, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 12.22, |
|
"grad_norm": 2.021256446838379, |
|
"learning_rate": 7.228235294117648e-05, |
|
"loss": 0.3293, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 12.22, |
|
"eval_loss": 0.29065316915512085, |
|
"eval_runtime": 168.1265, |
|
"eval_samples_per_second": 18.861, |
|
"eval_steps_per_second": 2.361, |
|
"eval_wer": 0.290982895604205, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 16.29, |
|
"grad_norm": 1.2028392553329468, |
|
"learning_rate": 6.443921568627452e-05, |
|
"loss": 0.203, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 16.29, |
|
"eval_loss": 0.29165971279144287, |
|
"eval_runtime": 168.0318, |
|
"eval_samples_per_second": 18.871, |
|
"eval_steps_per_second": 2.363, |
|
"eval_wer": 0.2812337369461888, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 20.37, |
|
"grad_norm": 0.8678894639015198, |
|
"learning_rate": 5.6596078431372547e-05, |
|
"loss": 0.1526, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 20.37, |
|
"eval_loss": 0.2595488429069519, |
|
"eval_runtime": 168.4253, |
|
"eval_samples_per_second": 18.827, |
|
"eval_steps_per_second": 2.357, |
|
"eval_wer": 0.26881310064878744, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 24.44, |
|
"grad_norm": 1.7510905265808105, |
|
"learning_rate": 4.875294117647059e-05, |
|
"loss": 0.1189, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 24.44, |
|
"eval_loss": 0.2896680533885956, |
|
"eval_runtime": 167.8147, |
|
"eval_samples_per_second": 18.896, |
|
"eval_steps_per_second": 2.366, |
|
"eval_wer": 0.2652395656246747, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 28.51, |
|
"grad_norm": 1.1846829652786255, |
|
"learning_rate": 4.090980392156863e-05, |
|
"loss": 0.0988, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 28.51, |
|
"eval_loss": 0.30395135283470154, |
|
"eval_runtime": 168.5254, |
|
"eval_samples_per_second": 18.816, |
|
"eval_steps_per_second": 2.356, |
|
"eval_wer": 0.2631925892516393, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 32.59, |
|
"grad_norm": 1.2768597602844238, |
|
"learning_rate": 3.3066666666666666e-05, |
|
"loss": 0.0767, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 32.59, |
|
"eval_loss": 0.2856810986995697, |
|
"eval_runtime": 170.6385, |
|
"eval_samples_per_second": 18.583, |
|
"eval_steps_per_second": 2.327, |
|
"eval_wer": 0.25576796308503624, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 36.66, |
|
"grad_norm": 0.8570753931999207, |
|
"learning_rate": 2.5223529411764707e-05, |
|
"loss": 0.0653, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 36.66, |
|
"eval_loss": 0.30118483304977417, |
|
"eval_runtime": 171.6141, |
|
"eval_samples_per_second": 18.478, |
|
"eval_steps_per_second": 2.313, |
|
"eval_wer": 0.24785761371127224, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 40.73, |
|
"grad_norm": 0.5461006164550781, |
|
"learning_rate": 1.7380392156862745e-05, |
|
"loss": 0.0585, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 40.73, |
|
"eval_loss": 0.28294435143470764, |
|
"eval_runtime": 169.3236, |
|
"eval_samples_per_second": 18.727, |
|
"eval_steps_per_second": 2.345, |
|
"eval_wer": 0.24251465843250183, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 44.81, |
|
"grad_norm": 0.47187602519989014, |
|
"learning_rate": 9.537254901960786e-06, |
|
"loss": 0.0467, |
|
"step": 5500 |
|
}, |
|
{ |
|
"epoch": 44.81, |
|
"eval_loss": 0.28480401635169983, |
|
"eval_runtime": 169.9442, |
|
"eval_samples_per_second": 18.659, |
|
"eval_steps_per_second": 2.336, |
|
"eval_wer": 0.23873295631960587, |
|
"step": 5500 |
|
}, |
|
{ |
|
"epoch": 48.88, |
|
"grad_norm": 0.4703833758831024, |
|
"learning_rate": 1.6941176470588237e-06, |
|
"loss": 0.0416, |
|
"step": 6000 |
|
}, |
|
{ |
|
"epoch": 48.88, |
|
"eval_loss": 0.27989715337753296, |
|
"eval_runtime": 170.4594, |
|
"eval_samples_per_second": 18.603, |
|
"eval_steps_per_second": 2.329, |
|
"eval_wer": 0.23436144745515733, |
|
"step": 6000 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 6100, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 50, |
|
"save_steps": 500, |
|
"total_flos": 9.05423649471038e+19, |
|
"train_batch_size": 8, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|