trainer_state.json

{
  "best_metric": 0.8207207322120667,
  "best_model_checkpoint": "model_llama2_7b_chat_hu_v2/checkpoint-18000",
  "epoch": 0.75,
  "eval_steps": 6000,
  "global_step": 18000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "learning_rate": 1.95e-05,
      "loss": 1.0674,
      "step": 6000
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.8339963555335999,
      "eval_runtime": 3670.1403,
      "eval_samples_per_second": 1.635,
      "eval_steps_per_second": 1.635,
      "step": 6000
    },
    {
      "epoch": 0.5,
      "learning_rate": 1.9e-05,
      "loss": 0.823,
      "step": 12000
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.8208308219909668,
      "eval_runtime": 3669.2437,
      "eval_samples_per_second": 1.635,
      "eval_steps_per_second": 1.635,
      "step": 12000
    },
    {
      "epoch": 0.75,
      "learning_rate": 1.8500000000000002e-05,
      "loss": 0.815,
      "step": 18000
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.8207207322120667,
      "eval_runtime": 3671.2164,
      "eval_samples_per_second": 1.634,
      "eval_steps_per_second": 1.634,
      "step": 18000
    }
  ],
  "logging_steps": 6000,
  "max_steps": 240000,
  "num_train_epochs": 10,
  "save_steps": 6000,
  "total_flos": 1.461438690361344e+18,
  "trial_name": null,
  "trial_params": null
}