xlmRoBERTa_polish / checkpoint-2772 /trainer_state.json
ringorsolya's picture
Upload folder using huggingface_hub
3a5a325 verified
{
"best_metric": 0.05848415940999985,
"best_model_checkpoint": "/content/drive/MyDrive/checkpoint-2772",
"epoch": 3.0,
"eval_steps": 100,
"global_step": 2772,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"grad_norm": 5.851963043212891,
"learning_rate": 4.5e-06,
"loss": 0.261,
"step": 924
},
{
"epoch": 1.0,
"eval_accuracy": 0.9312026002166848,
"eval_f1": 0.9304556698132427,
"eval_loss": 0.10791309922933578,
"eval_precision": 0.9337159575536489,
"eval_recall": 0.9312026002166848,
"eval_runtime": 49.9811,
"eval_samples_per_second": 36.934,
"eval_steps_per_second": 2.321,
"step": 924
},
{
"epoch": 2.0,
"grad_norm": 0.24086147546768188,
"learning_rate": 4.000000000000001e-06,
"loss": 0.0812,
"step": 1848
},
{
"epoch": 2.0,
"eval_accuracy": 0.9582881906825569,
"eval_f1": 0.9581837801589876,
"eval_loss": 0.06059493124485016,
"eval_precision": 0.9585273431262236,
"eval_recall": 0.9582881906825569,
"eval_runtime": 49.6996,
"eval_samples_per_second": 37.143,
"eval_steps_per_second": 2.334,
"step": 1848
},
{
"epoch": 3.0,
"grad_norm": 0.1482285112142563,
"learning_rate": 3.5e-06,
"loss": 0.0501,
"step": 2772
},
{
"epoch": 3.0,
"eval_accuracy": 0.9544962080173348,
"eval_f1": 0.9543281540493387,
"eval_loss": 0.05848415940999985,
"eval_precision": 0.9552102707240017,
"eval_recall": 0.9544962080173348,
"eval_runtime": 49.6249,
"eval_samples_per_second": 37.199,
"eval_steps_per_second": 2.338,
"step": 2772
}
],
"logging_steps": 100,
"max_steps": 9240,
"num_input_tokens_seen": 0,
"num_train_epochs": 10,
"save_steps": 100,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 2,
"early_stopping_threshold": 0.0
},
"attributes": {
"early_stopping_patience_counter": 0
}
},
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 1.165886956910592e+16,
"train_batch_size": 16,
"trial_name": null,
"trial_params": null
}