|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 48.379746835443036, |
|
"global_step": 1500, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 8.06, |
|
"learning_rate": 0.0005, |
|
"loss": 0.576, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 8.06, |
|
"eval_cer": 0.050238462365976976, |
|
"eval_loss": 0.24107621610164642, |
|
"eval_runtime": 380.2463, |
|
"eval_samples_per_second": 19.111, |
|
"eval_steps_per_second": 2.391, |
|
"eval_wer": 0.23400590063956597, |
|
"step": 250 |
|
}, |
|
{ |
|
"epoch": 16.13, |
|
"learning_rate": 0.001, |
|
"loss": 0.2564, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 16.13, |
|
"eval_cer": 0.04923441029870549, |
|
"eval_loss": 0.23052524030208588, |
|
"eval_runtime": 267.8454, |
|
"eval_samples_per_second": 27.131, |
|
"eval_steps_per_second": 3.394, |
|
"eval_wer": 0.20967071263093282, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 24.19, |
|
"learning_rate": 0.0007619047619047619, |
|
"loss": 0.2018, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 24.19, |
|
"eval_cer": 0.04937458639900638, |
|
"eval_loss": 0.23705527186393738, |
|
"eval_runtime": 264.0504, |
|
"eval_samples_per_second": 27.521, |
|
"eval_steps_per_second": 3.443, |
|
"eval_wer": 0.20590855988753143, |
|
"step": 750 |
|
}, |
|
{ |
|
"epoch": 32.25, |
|
"learning_rate": 0.0005238095238095238, |
|
"loss": 0.1549, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 32.25, |
|
"eval_cer": 0.04349696991429698, |
|
"eval_loss": 0.22975854575634003, |
|
"eval_runtime": 264.3572, |
|
"eval_samples_per_second": 27.489, |
|
"eval_steps_per_second": 3.439, |
|
"eval_wer": 0.18440488683840564, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 40.32, |
|
"learning_rate": 0.0002857142857142857, |
|
"loss": 0.1224, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 40.32, |
|
"eval_cer": 0.040729306910681745, |
|
"eval_loss": 0.22875599563121796, |
|
"eval_runtime": 265.748, |
|
"eval_samples_per_second": 27.345, |
|
"eval_steps_per_second": 3.421, |
|
"eval_wer": 0.1724650020790844, |
|
"step": 1250 |
|
}, |
|
{ |
|
"epoch": 48.38, |
|
"learning_rate": 4.761904761904762e-05, |
|
"loss": 0.1004, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 48.38, |
|
"eval_cer": 0.03763565297613421, |
|
"eval_loss": 0.23267094790935516, |
|
"eval_runtime": 266.4102, |
|
"eval_samples_per_second": 27.277, |
|
"eval_steps_per_second": 3.412, |
|
"eval_wer": 0.16082212937845278, |
|
"step": 1500 |
|
} |
|
], |
|
"max_steps": 1550, |
|
"num_train_epochs": 50, |
|
"total_flos": 1.244071149131343e+20, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|