trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 3920,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.64,
      "learning_rate": 2.6173469387755103e-05,
      "loss": 0.0854,
      "step": 500
    },
    {
      "epoch": 1.28,
      "learning_rate": 2.2346938775510205e-05,
      "loss": 0.0409,
      "step": 1000
    },
    {
      "epoch": 1.91,
      "learning_rate": 1.8520408163265304e-05,
      "loss": 0.0289,
      "step": 1500
    },
    {
      "epoch": 2.55,
      "learning_rate": 1.4693877551020408e-05,
      "loss": 0.0187,
      "step": 2000
    },
    {
      "epoch": 3.19,
      "learning_rate": 1.086734693877551e-05,
      "loss": 0.0147,
      "step": 2500
    },
    {
      "epoch": 3.83,
      "learning_rate": 7.0408163265306125e-06,
      "loss": 0.0083,
      "step": 3000
    },
    {
      "epoch": 4.46,
      "learning_rate": 3.2142857142857143e-06,
      "loss": 0.0056,
      "step": 3500
    },
    {
      "epoch": 5.0,
      "step": 3920,
      "total_flos": 6628339113716736.0,
      "train_loss": 0.026263209812495174,
      "train_runtime": 785.6494,
      "train_samples_per_second": 79.832,
      "train_steps_per_second": 4.99
    }
  ],
  "max_steps": 3920,
  "num_train_epochs": 5,
  "total_flos": 6628339113716736.0,
  "trial_name": null,
  "trial_params": null
}