trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 1136,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 4.995598591549296e-05,
      "loss": 2.3044,
      "step": 1
    },
    {
      "epoch": 3.52,
      "learning_rate": 2.7992957746478874e-05,
      "loss": 0.9153,
      "step": 500
    },
    {
      "epoch": 7.04,
      "learning_rate": 6.0299295774647895e-06,
      "loss": 0.0323,
      "step": 1000
    },
    {
      "epoch": 8.0,
      "step": 1136,
      "total_flos": 2.844093826895708e+18,
      "train_loss": 0.4185669529973201,
      "train_runtime": 25964.0811,
      "train_samples_per_second": 0.7,
      "train_steps_per_second": 0.044
    }
  ],
  "logging_steps": 500,
  "max_steps": 1136,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 2.844093826895708e+18,
  "trial_name": null,
  "trial_params": null
}