last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.0019779133014669525,
  "eval_steps": 3,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0006593044338223175,
      "grad_norm": 14.806910514831543,
      "learning_rate": 2e-05,
      "loss": 8.7677,
      "step": 1
    },
    {
      "epoch": 0.001318608867644635,
      "grad_norm": 18.260412216186523,
      "learning_rate": 4e-05,
      "loss": 8.3319,
      "step": 2
    },
    {
      "epoch": 0.0019779133014669525,
      "grad_norm": 14.558311462402344,
      "learning_rate": 6e-05,
      "loss": 8.6052,
      "step": 3
    },
    {
      "epoch": 0.0019779133014669525,
      "eval_loss": 2.1339499950408936,
      "eval_runtime": 38.3719,
      "eval_samples_per_second": 16.653,
      "eval_steps_per_second": 8.339,
      "step": 3
    }
  ],
  "logging_steps": 1,
  "max_steps": 10,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 3,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 525800935784448.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}