trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.6,
  "eval_steps": 500,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 51.644339588858706,
      "learning_rate": 5e-06,
      "loss": 1.9079,
      "step": 1
    },
    {
      "epoch": 0.5333333333333333,
      "eval_loss": 1.8345922231674194,
      "eval_runtime": 13.5887,
      "eval_samples_per_second": 7.433,
      "eval_steps_per_second": 0.294,
      "step": 1
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 36.69583629812056,
      "learning_rate": 5e-06,
      "loss": 1.8231,
      "step": 2
    },
    {
      "epoch": 1.6,
      "grad_norm": 121.54965597378022,
      "learning_rate": 5e-06,
      "loss": 1.7235,
      "step": 3
    },
    {
      "epoch": 1.6,
      "eval_loss": 1.7085474729537964,
      "eval_runtime": 13.4334,
      "eval_samples_per_second": 7.519,
      "eval_steps_per_second": 0.298,
      "step": 3
    },
    {
      "epoch": 1.6,
      "step": 3,
      "total_flos": 1.922102501100749e+16,
      "train_loss": 1.818156321843465,
      "train_runtime": 1285.2745,
      "train_samples_per_second": 4.479,
      "train_steps_per_second": 0.002
    }
  ],
  "logging_steps": 1,
  "max_steps": 3,
  "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": 1.922102501100749e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}