checkpoint-21/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.4285714285714286,
      "grad_norm": 2.016594171524048,
      "learning_rate": 2.6190476190476192e-05,
      "loss": 1.3059,
      "step": 10
    },
    {
      "epoch": 2.857142857142857,
      "grad_norm": 2.60115385055542,
      "learning_rate": 2.3809523809523808e-06,
      "loss": 1.1869,
      "step": 20
    }
  ],
  "logging_steps": 10,
  "max_steps": 21,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 682174488060.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}