checkpoint-50/trainer_state.json

{
  "best_metric": 1.083932638168335,
  "best_model_checkpoint": "checkpoints/7b/checkpoint-50",
  "epoch": 0.08800880088008801,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04,
      "learning_rate": 2.8499999999999998e-05,
      "loss": 2.0006,
      "step": 20
    },
    {
      "epoch": 0.07,
      "learning_rate": 5.85e-05,
      "loss": 1.6054,
      "step": 40
    },
    {
      "epoch": 0.09,
      "eval_loss": 1.083932638168335,
      "eval_runtime": 1647.9375,
      "eval_samples_per_second": 9.456,
      "eval_steps_per_second": 1.182,
      "step": 50
    }
  ],
  "max_steps": 568,
  "num_train_epochs": 1,
  "total_flos": 6.52559141044224e+16,
  "trial_name": null,
  "trial_params": null
}