|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 511.7692307692308, |
|
"global_step": 3071, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 16.62, |
|
"learning_rate": 2.4451954397394137e-05, |
|
"loss": 14.5899, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 33.31, |
|
"learning_rate": 4.863762214983713e-05, |
|
"loss": 5.5209, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 49.92, |
|
"learning_rate": 7.282328990228012e-05, |
|
"loss": 3.8166, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 66.62, |
|
"learning_rate": 7.5e-05, |
|
"loss": 3.5875, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 66.62, |
|
"eval_cer": 1.0, |
|
"eval_loss": 3.4289283752441406, |
|
"eval_runtime": 10.7295, |
|
"eval_samples_per_second": 20.318, |
|
"eval_steps_per_second": 0.373, |
|
"eval_wer": 1.0, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 83.31, |
|
"learning_rate": 7.5e-05, |
|
"loss": 3.3325, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 99.92, |
|
"learning_rate": 7.5e-05, |
|
"loss": 2.1768, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 116.62, |
|
"learning_rate": 7.5e-05, |
|
"loss": 1.6244, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 133.31, |
|
"learning_rate": 7.5e-05, |
|
"loss": 1.4065, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 133.31, |
|
"eval_cer": 0.1734040203225094, |
|
"eval_loss": 0.7243412733078003, |
|
"eval_runtime": 10.6388, |
|
"eval_samples_per_second": 20.491, |
|
"eval_steps_per_second": 0.376, |
|
"eval_wer": 0.6619263089851325, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 149.92, |
|
"learning_rate": 7.5e-05, |
|
"loss": 1.2615, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 166.62, |
|
"learning_rate": 7.5e-05, |
|
"loss": 1.1668, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 183.31, |
|
"learning_rate": 7.5e-05, |
|
"loss": 1.0808, |
|
"step": 1100 |
|
}, |
|
{ |
|
"epoch": 199.92, |
|
"learning_rate": 7.5e-05, |
|
"loss": 1.007, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 199.92, |
|
"eval_cer": 0.1456814667550254, |
|
"eval_loss": 0.6611430048942566, |
|
"eval_runtime": 10.4906, |
|
"eval_samples_per_second": 20.78, |
|
"eval_steps_per_second": 0.381, |
|
"eval_wer": 0.5830639948287007, |
|
"step": 1200 |
|
}, |
|
{ |
|
"epoch": 216.62, |
|
"learning_rate": 7.5e-05, |
|
"loss": 0.9543, |
|
"step": 1300 |
|
}, |
|
{ |
|
"epoch": 233.31, |
|
"learning_rate": 7.5e-05, |
|
"loss": 0.8939, |
|
"step": 1400 |
|
}, |
|
{ |
|
"epoch": 249.92, |
|
"learning_rate": 7.5e-05, |
|
"loss": 0.842, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 266.62, |
|
"learning_rate": 7.207763671875e-05, |
|
"loss": 0.7984, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 266.62, |
|
"eval_cer": 0.13320079522862824, |
|
"eval_loss": 0.6387308239936829, |
|
"eval_runtime": 10.4188, |
|
"eval_samples_per_second": 20.924, |
|
"eval_steps_per_second": 0.384, |
|
"eval_wer": 0.5520361990950227, |
|
"step": 1600 |
|
}, |
|
{ |
|
"epoch": 283.31, |
|
"learning_rate": 6.743896484375e-05, |
|
"loss": 0.7432, |
|
"step": 1700 |
|
}, |
|
{ |
|
"epoch": 299.92, |
|
"learning_rate": 6.280029296875e-05, |
|
"loss": 0.6906, |
|
"step": 1800 |
|
}, |
|
{ |
|
"epoch": 316.62, |
|
"learning_rate": 5.816162109375e-05, |
|
"loss": 0.6542, |
|
"step": 1900 |
|
}, |
|
{ |
|
"epoch": 333.31, |
|
"learning_rate": 5.3522949218749995e-05, |
|
"loss": 0.6117, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 333.31, |
|
"eval_cer": 0.13761873205213165, |
|
"eval_loss": 0.7423861622810364, |
|
"eval_runtime": 10.4252, |
|
"eval_samples_per_second": 20.911, |
|
"eval_steps_per_second": 0.384, |
|
"eval_wer": 0.5681965093729799, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 349.92, |
|
"learning_rate": 4.888427734375e-05, |
|
"loss": 0.5816, |
|
"step": 2100 |
|
}, |
|
{ |
|
"epoch": 366.62, |
|
"learning_rate": 4.424560546875e-05, |
|
"loss": 0.5444, |
|
"step": 2200 |
|
}, |
|
{ |
|
"epoch": 383.31, |
|
"learning_rate": 3.9606933593749996e-05, |
|
"loss": 0.5214, |
|
"step": 2300 |
|
}, |
|
{ |
|
"epoch": 399.92, |
|
"learning_rate": 3.50146484375e-05, |
|
"loss": 0.4926, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 399.92, |
|
"eval_cer": 0.13143362049922686, |
|
"eval_loss": 0.7627450823783875, |
|
"eval_runtime": 10.7581, |
|
"eval_samples_per_second": 20.264, |
|
"eval_steps_per_second": 0.372, |
|
"eval_wer": 0.5513897866839044, |
|
"step": 2400 |
|
}, |
|
{ |
|
"epoch": 416.62, |
|
"learning_rate": 3.0375976562499997e-05, |
|
"loss": 0.4671, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 433.31, |
|
"learning_rate": 2.5737304687499997e-05, |
|
"loss": 0.445, |
|
"step": 2600 |
|
}, |
|
{ |
|
"epoch": 449.92, |
|
"learning_rate": 2.1098632812500003e-05, |
|
"loss": 0.4303, |
|
"step": 2700 |
|
}, |
|
{ |
|
"epoch": 466.62, |
|
"learning_rate": 1.64599609375e-05, |
|
"loss": 0.416, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 466.62, |
|
"eval_cer": 0.1319858626021648, |
|
"eval_loss": 0.7815532088279724, |
|
"eval_runtime": 10.5614, |
|
"eval_samples_per_second": 20.641, |
|
"eval_steps_per_second": 0.379, |
|
"eval_wer": 0.5604395604395604, |
|
"step": 2800 |
|
}, |
|
{ |
|
"epoch": 483.31, |
|
"learning_rate": 1.1821289062500006e-05, |
|
"loss": 0.3988, |
|
"step": 2900 |
|
}, |
|
{ |
|
"epoch": 499.92, |
|
"learning_rate": 7.182617187500003e-06, |
|
"loss": 0.3935, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 511.77, |
|
"step": 3071, |
|
"total_flos": 8.928364517399495e+19, |
|
"train_loss": 1.6842556819586596, |
|
"train_runtime": 26421.1393, |
|
"train_samples_per_second": 14.878, |
|
"train_steps_per_second": 0.116 |
|
} |
|
], |
|
"max_steps": 3071, |
|
"num_train_epochs": 512, |
|
"total_flos": 8.928364517399495e+19, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|