trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.010057471264367816,
  "global_step": 77,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.4805194805194807e-05,
      "loss": 1.4342,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 9.610389610389611e-06,
      "loss": 1.3017,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 4.415584415584416e-06,
      "loss": 1.3118,
      "step": 60
    },
    {
      "epoch": 0.01,
      "step": 77,
      "total_flos": 28330426368.0,
      "train_loss": 1.2590427150974026,
      "train_runtime": 57.7633,
      "train_samples_per_second": 1.325,
      "train_steps_per_second": 1.333
    }
  ],
  "max_steps": 77,
  "num_train_epochs": 1,
  "total_flos": 28330426368.0,
  "trial_name": null,
  "trial_params": null
}