trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.8888888888888888,
  "eval_steps": 500,
  "global_step": 7,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12698412698412698,
      "grad_norm": 1.2362275044985542,
      "learning_rate": 0.0,
      "loss": 1.314,
      "step": 1
    },
    {
      "epoch": 0.25396825396825395,
      "grad_norm": 0.976147403234752,
      "learning_rate": 0.0002,
      "loss": 1.2319,
      "step": 2
    },
    {
      "epoch": 0.38095238095238093,
      "grad_norm": 0.9870148336093962,
      "learning_rate": 0.0002,
      "loss": 1.1567,
      "step": 3
    },
    {
      "epoch": 0.5079365079365079,
      "grad_norm": 0.5333773988410998,
      "learning_rate": 0.0002,
      "loss": 1.0859,
      "step": 4
    },
    {
      "epoch": 0.6349206349206349,
      "grad_norm": 0.3508843450976753,
      "learning_rate": 0.0002,
      "loss": 0.9665,
      "step": 5
    },
    {
      "epoch": 0.7619047619047619,
      "grad_norm": 0.46075909604963405,
      "learning_rate": 0.0002,
      "loss": 0.9759,
      "step": 6
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 0.25684210186076895,
      "learning_rate": 0.0002,
      "loss": 0.9419,
      "step": 7
    },
    {
      "epoch": 0.8888888888888888,
      "step": 7,
      "total_flos": 14802352734208.0,
      "train_loss": 1.0961060779435294,
      "train_runtime": 3701.1484,
      "train_samples_per_second": 0.272,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 7,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 14802352734208.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}