0x1202's picture
Training in progress, step 1, checkpoint
adcc5b0 verified
{
"best_metric": 10.793596267700195,
"best_model_checkpoint": "miner_id_24/checkpoint-1",
"epoch": 1.0,
"eval_steps": 50,
"global_step": 1,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"grad_norm": 2.5258162021636963,
"learning_rate": 1e-05,
"loss": 10.6314,
"step": 1
},
{
"epoch": 1.0,
"eval_loss": 10.793596267700195,
"eval_runtime": 0.0341,
"eval_samples_per_second": 29.298,
"eval_steps_per_second": 29.298,
"step": 1
}
],
"logging_steps": 1,
"max_steps": 1,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 50,
"stateful_callbacks": {
"EarlyStoppingCallback": {
"args": {
"early_stopping_patience": 5,
"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": 5636087611392.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}