trainer_state.json

{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 5.0,
      "grad_norm": 0.09996598958969116,
      "learning_rate": 0.00019994755690455152,
      "loss": 0.2824,
      "step": 5
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.056030746549367905,
      "learning_rate": 0.0001981178176898239,
      "loss": 0.1544,
      "step": 10
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.031677622348070145,
      "learning_rate": 0.00019372067050063438,
      "loss": 0.1034,
      "step": 15
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.017535891383886337,
      "learning_rate": 0.00018687117365181512,
      "loss": 0.0666,
      "step": 20
    }
  ],
  "logging_steps": 5,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 100,
  "save_steps": 20,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3.1802671723708416e+17,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}