trainer_state.json

{
  "best_metric": 0.3832016885280609,
  "best_model_checkpoint": "./lora-html/checkpoint-100",
  "epoch": 5.2631578947368425,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.63,
      "learning_rate": 0.00014399999999999998,
      "loss": 1.0623,
      "step": 50
    },
    {
      "epoch": 2.63,
      "eval_loss": 0.5735694766044617,
      "eval_runtime": 2.0662,
      "eval_samples_per_second": 9.68,
      "eval_steps_per_second": 1.452,
      "step": 50
    },
    {
      "epoch": 5.26,
      "learning_rate": 0.000294,
      "loss": 0.4702,
      "step": 100
    },
    {
      "epoch": 5.26,
      "eval_loss": 0.3832016885280609,
      "eval_runtime": 2.2136,
      "eval_samples_per_second": 9.035,
      "eval_steps_per_second": 1.355,
      "step": 100
    }
  ],
  "max_steps": 100,
  "num_train_epochs": 6,
  "total_flos": 3.135836567764992e+16,
  "trial_name": null,
  "trial_params": null
}