trainer_state.json

{
  "best_metric": 1.0345826148986816,
  "best_model_checkpoint": "Llama-2_7B-chat_3/checkpoint-30",
  "epoch": 1.9186602870813396,
  "eval_steps": 10,
  "global_step": 30,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.77,
      "eval_loss": 1.2624627351760864,
      "eval_runtime": 67.2768,
      "eval_samples_per_second": 6.213,
      "eval_steps_per_second": 0.401,
      "step": 10
    },
    {
      "epoch": 1.15,
      "eval_loss": 1.1051764488220215,
      "eval_runtime": 66.3159,
      "eval_samples_per_second": 6.303,
      "eval_steps_per_second": 0.407,
      "step": 20
    },
    {
      "epoch": 1.92,
      "eval_loss": 1.0345826148986816,
      "eval_runtime": 66.399,
      "eval_samples_per_second": 6.295,
      "eval_steps_per_second": 0.407,
      "step": 30
    }
  ],
  "logging_steps": 40,
  "max_steps": 39,
  "num_train_epochs": 3,
  "save_steps": 10,
  "total_flos": 1.564336319449006e+17,
  "trial_name": null,
  "trial_params": null
}