|
{ |
|
"best_metric": 0.7239218354225159, |
|
"best_model_checkpoint": "distilbert_lda_50_v1_wnli/checkpoint-9", |
|
"epoch": 8.0, |
|
"eval_steps": 500, |
|
"global_step": 24, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 2.593956708908081, |
|
"learning_rate": 4.9e-05, |
|
"loss": 0.7829, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.3380281690140845, |
|
"eval_loss": 0.7284329533576965, |
|
"eval_runtime": 0.0557, |
|
"eval_samples_per_second": 1274.439, |
|
"eval_steps_per_second": 17.95, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 1.2626270055770874, |
|
"learning_rate": 4.8e-05, |
|
"loss": 0.7048, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.3380281690140845, |
|
"eval_loss": 0.7267274260520935, |
|
"eval_runtime": 0.038, |
|
"eval_samples_per_second": 1867.408, |
|
"eval_steps_per_second": 26.302, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 1.094574213027954, |
|
"learning_rate": 4.7e-05, |
|
"loss": 0.689, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.49295774647887325, |
|
"eval_loss": 0.7239218354225159, |
|
"eval_runtime": 0.0391, |
|
"eval_samples_per_second": 1815.749, |
|
"eval_steps_per_second": 25.574, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 1.0490866899490356, |
|
"learning_rate": 4.600000000000001e-05, |
|
"loss": 0.6915, |
|
"step": 12 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.39436619718309857, |
|
"eval_loss": 0.7566296458244324, |
|
"eval_runtime": 0.0395, |
|
"eval_samples_per_second": 1797.143, |
|
"eval_steps_per_second": 25.312, |
|
"step": 12 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 1.0118927955627441, |
|
"learning_rate": 4.5e-05, |
|
"loss": 0.6916, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.2535211267605634, |
|
"eval_loss": 0.754951536655426, |
|
"eval_runtime": 0.0404, |
|
"eval_samples_per_second": 1756.274, |
|
"eval_steps_per_second": 24.736, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 1.1331769227981567, |
|
"learning_rate": 4.4000000000000006e-05, |
|
"loss": 0.6936, |
|
"step": 18 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.2535211267605634, |
|
"eval_loss": 0.7512102127075195, |
|
"eval_runtime": 0.0378, |
|
"eval_samples_per_second": 1876.634, |
|
"eval_steps_per_second": 26.431, |
|
"step": 18 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"grad_norm": 0.7647018432617188, |
|
"learning_rate": 4.3e-05, |
|
"loss": 0.6908, |
|
"step": 21 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.2535211267605634, |
|
"eval_loss": 0.7626540064811707, |
|
"eval_runtime": 0.0417, |
|
"eval_samples_per_second": 1701.864, |
|
"eval_steps_per_second": 23.97, |
|
"step": 21 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 2.0840747356414795, |
|
"learning_rate": 4.2e-05, |
|
"loss": 0.6841, |
|
"step": 24 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.18309859154929578, |
|
"eval_loss": 0.7832306623458862, |
|
"eval_runtime": 0.0385, |
|
"eval_samples_per_second": 1843.148, |
|
"eval_steps_per_second": 25.96, |
|
"step": 24 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"step": 24, |
|
"total_flos": 336467192586240.0, |
|
"train_loss": 0.7035509645938873, |
|
"train_runtime": 19.7006, |
|
"train_samples_per_second": 1611.629, |
|
"train_steps_per_second": 7.614 |
|
} |
|
], |
|
"logging_steps": 1, |
|
"max_steps": 150, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 50, |
|
"save_steps": 500, |
|
"stateful_callbacks": { |
|
"EarlyStoppingCallback": { |
|
"args": { |
|
"early_stopping_patience": 5, |
|
"early_stopping_threshold": 0.0 |
|
}, |
|
"attributes": { |
|
"early_stopping_patience_counter": 5 |
|
} |
|
}, |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": true |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 336467192586240.0, |
|
"train_batch_size": 256, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|