trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.028011204481792718,
  "eval_steps": 50,
  "global_step": 5,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.5943,
      "step": 1
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.00017071067811865476,
      "loss": 0.4897,
      "step": 2
    },
    {
      "epoch": 0.02,
      "learning_rate": 2.9289321881345254e-05,
      "loss": 0.3496,
      "step": 4
    },
    {
      "epoch": 0.03,
      "step": 5,
      "total_flos": 2.119653912064819e+16,
      "train_loss": 0.4382333517074585,
      "train_runtime": 136.4197,
      "train_samples_per_second": 0.586,
      "train_steps_per_second": 0.037
    }
  ],
  "logging_steps": 2,
  "max_steps": 5,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 200,
  "total_flos": 2.119653912064819e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}