trainer_state.json

{
  "best_global_step": 50,
  "best_metric": 1.7122082710266113,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 25,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 5.170515537261963,
      "learning_rate": 9.444444444444445e-06,
      "loss": 3.801,
      "step": 10
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 6.68876314163208,
      "learning_rate": 7.592592592592594e-06,
      "loss": 3.4564,
      "step": 20
    },
    {
      "epoch": 2.0833333333333335,
      "eval_loss": 1.7632215023040771,
      "eval_runtime": 0.9476,
      "eval_samples_per_second": 86.536,
      "eval_steps_per_second": 6.332,
      "step": 25
    },
    {
      "epoch": 2.5,
      "grad_norm": 6.427336692810059,
      "learning_rate": 5.740740740740741e-06,
      "loss": 3.2725,
      "step": 30
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 7.084609508514404,
      "learning_rate": 3.88888888888889e-06,
      "loss": 3.152,
      "step": 40
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 7.747525215148926,
      "learning_rate": 2.037037037037037e-06,
      "loss": 3.0422,
      "step": 50
    },
    {
      "epoch": 4.166666666666667,
      "eval_loss": 1.7122082710266113,
      "eval_runtime": 0.9379,
      "eval_samples_per_second": 87.432,
      "eval_steps_per_second": 6.397,
      "step": 50
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.486431121826172,
      "learning_rate": 1.8518518518518518e-07,
      "loss": 2.9745,
      "step": 60
    }
  ],
  "logging_steps": 10,
  "max_steps": 60,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 300,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}