trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.571428571428571,
  "eval_steps": 500,
  "global_step": 9,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2857142857142857,
      "grad_norm": 16229.822717765965,
      "learning_rate": 2e-05,
      "loss": 63.281,
      "step": 1
    },
    {
      "epoch": 0.8571428571428571,
      "eval_loss": 31.52655601501465,
      "eval_runtime": 0.8665,
      "eval_samples_per_second": 24.235,
      "eval_steps_per_second": 1.154,
      "step": 3
    },
    {
      "epoch": 1.4285714285714286,
      "grad_norm": 20842.87155371142,
      "learning_rate": 1e-05,
      "loss": 53.4745,
      "step": 5
    },
    {
      "epoch": 2.0,
      "eval_loss": 24.942411422729492,
      "eval_runtime": 0.841,
      "eval_samples_per_second": 24.971,
      "eval_steps_per_second": 1.189,
      "step": 7
    },
    {
      "epoch": 2.571428571428571,
      "eval_loss": 24.173879623413086,
      "eval_runtime": 0.8315,
      "eval_samples_per_second": 25.255,
      "eval_steps_per_second": 1.203,
      "step": 9
    },
    {
      "epoch": 2.571428571428571,
      "step": 9,
      "total_flos": 2409879306240.0,
      "train_loss": 42.46108203464084,
      "train_runtime": 161.2984,
      "train_samples_per_second": 7.812,
      "train_steps_per_second": 0.056
    }
  ],
  "logging_steps": 5,
  "max_steps": 9,
  "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": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2409879306240.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}