trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 1875,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.33,
      "learning_rate": 3.6666666666666666e-05,
      "loss": 0.8661,
      "step": 500
    },
    {
      "epoch": 2.67,
      "learning_rate": 2.3333333333333336e-05,
      "loss": 0.0521,
      "step": 1000
    },
    {
      "epoch": 4.0,
      "learning_rate": 1e-05,
      "loss": 0.0199,
      "step": 1500
    },
    {
      "epoch": 5.0,
      "step": 1875,
      "total_flos": 31746111307776.0,
      "train_loss": 0.2530242416381836,
      "train_runtime": 134.3349,
      "train_samples_per_second": 55.831,
      "train_steps_per_second": 13.958
    }
  ],
  "max_steps": 1875,
  "num_train_epochs": 5,
  "total_flos": 31746111307776.0,
  "trial_name": null,
  "trial_params": null
}