syn-news-dict / trainer_state.json
layai's picture
Add files using upload-large-folder tool
04d8051 verified
{
"best_metric": 1.9098191261291504,
"best_model_checkpoint": "./news/dict/checkpoint-500",
"epoch": 2.996832101372756,
"eval_steps": 500,
"global_step": 1419,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0559662090813093,
"grad_norm": 3.65625,
"learning_rate": 3.6184243046482264e-05,
"loss": 0.7964,
"step": 500
},
{
"epoch": 1.0559662090813093,
"eval_accuracy": 0.7301090501762594,
"eval_loss": 1.9098191261291504,
"eval_runtime": 3.8952,
"eval_samples_per_second": 61.358,
"eval_steps_per_second": 1.54,
"step": 500
},
{
"epoch": 2.1119324181626187,
"grad_norm": 2.953125,
"learning_rate": 1.0006983401822941e-05,
"loss": 0.2989,
"step": 1000
},
{
"epoch": 2.1119324181626187,
"eval_accuracy": 0.7473396369386881,
"eval_loss": 1.923977017402649,
"eval_runtime": 1.474,
"eval_samples_per_second": 162.142,
"eval_steps_per_second": 4.07,
"step": 1000
},
{
"epoch": 2.996832101372756,
"step": 1419,
"total_flos": 3.408124384169165e+16,
"train_loss": 0.4320870924365613,
"train_runtime": 2161.8514,
"train_samples_per_second": 2.628,
"train_steps_per_second": 0.656
}
],
"logging_steps": 500,
"max_steps": 1419,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 3,
"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": true
},
"attributes": {}
}
},
"total_flos": 3.408124384169165e+16,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}