trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 28,
  "global_step": 67,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "grad_norm": 9.4375,
      "learning_rate": 1.8153846153846155e-05,
      "loss": 0.4282,
      "step": 8
    },
    {
      "epoch": 0.24,
      "grad_norm": 10.5,
      "learning_rate": 1.5692307692307693e-05,
      "loss": 0.4335,
      "step": 16
    },
    {
      "epoch": 0.36,
      "grad_norm": 5.03125,
      "learning_rate": 1.3230769230769231e-05,
      "loss": 0.3984,
      "step": 24
    },
    {
      "epoch": 0.42,
      "eval_accuracy": 0.8166666666666667,
      "eval_f1_score": 0.8080459770114942,
      "eval_gmean": 0.7501820727030913,
      "eval_loss": 0.5816406011581421,
      "eval_precision": 0.8133022774327122,
      "eval_recall": 0.8166666666666667,
      "eval_runtime": 175.4354,
      "eval_samples_per_second": 0.342,
      "eval_steps_per_second": 0.046,
      "step": 28
    },
    {
      "epoch": 0.48,
      "grad_norm": 6.625,
      "learning_rate": 1.076923076923077e-05,
      "loss": 0.3979,
      "step": 32
    },
    {
      "epoch": 0.6,
      "grad_norm": 7.875,
      "learning_rate": 8.307692307692309e-06,
      "loss": 0.3657,
      "step": 40
    },
    {
      "epoch": 0.72,
      "grad_norm": 10.0,
      "learning_rate": 5.846153846153847e-06,
      "loss": 0.38,
      "step": 48
    },
    {
      "epoch": 0.84,
      "grad_norm": 9.6875,
      "learning_rate": 3.384615384615385e-06,
      "loss": 0.4054,
      "step": 56
    },
    {
      "epoch": 0.84,
      "eval_accuracy": 0.8166666666666667,
      "eval_f1_score": 0.8080459770114942,
      "eval_gmean": 0.7501820727030913,
      "eval_loss": 0.5850911736488342,
      "eval_precision": 0.8133022774327122,
      "eval_recall": 0.8166666666666667,
      "eval_runtime": 173.49,
      "eval_samples_per_second": 0.346,
      "eval_steps_per_second": 0.046,
      "step": 56
    },
    {
      "epoch": 0.96,
      "grad_norm": 8.8125,
      "learning_rate": 9.230769230769232e-07,
      "loss": 0.3864,
      "step": 64
    },
    {
      "epoch": 1.0,
      "step": 67,
      "total_flos": 8.435821078904832e+16,
      "train_loss": 0.405215135261194,
      "train_runtime": 12957.5785,
      "train_samples_per_second": 0.33,
      "train_steps_per_second": 0.005
    }
  ],
  "logging_steps": 8,
  "max_steps": 67,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 8.435821078904832e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}