trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 0.0,
      "loss": 18.7924,
      "step": 1
    },
    {
      "epoch": 2.0,
      "learning_rate": 0.0,
      "loss": 20.7055,
      "step": 2
    },
    {
      "epoch": 3.0,
      "learning_rate": 0.0,
      "loss": 20.7584,
      "step": 3
    },
    {
      "epoch": 4.0,
      "learning_rate": 0.0,
      "loss": 19.0565,
      "step": 4
    },
    {
      "epoch": 5.0,
      "learning_rate": 6e-07,
      "loss": 19.4821,
      "step": 5
    },
    {
      "epoch": 6.0,
      "learning_rate": 1.2e-06,
      "loss": 19.2156,
      "step": 6
    },
    {
      "epoch": 7.0,
      "learning_rate": 1.8e-06,
      "loss": 18.2756,
      "step": 7
    },
    {
      "epoch": 8.0,
      "learning_rate": 2.4e-06,
      "loss": 20.6991,
      "step": 8
    },
    {
      "epoch": 9.0,
      "learning_rate": 2.9999999999999997e-06,
      "loss": 16.1652,
      "step": 9
    },
    {
      "epoch": 10.0,
      "learning_rate": 3.6e-06,
      "loss": 19.7457,
      "step": 10
    },
    {
      "epoch": 11.0,
      "learning_rate": 4.2e-06,
      "loss": 19.2072,
      "step": 11
    },
    {
      "epoch": 12.0,
      "learning_rate": 4.8e-06,
      "loss": 16.8633,
      "step": 12
    },
    {
      "epoch": 13.0,
      "learning_rate": 5.399999999999999e-06,
      "loss": 19.2082,
      "step": 13
    },
    {
      "epoch": 14.0,
      "learning_rate": 5.999999999999999e-06,
      "loss": 20.4279,
      "step": 14
    },
    {
      "epoch": 15.0,
      "learning_rate": 6.599999999999999e-06,
      "loss": 20.6761,
      "step": 15
    },
    {
      "epoch": 15.0,
      "step": 15,
      "total_flos": 9.481946310220186e+16,
      "train_loss": 19.285240809122723,
      "train_runtime": 17.2618,
      "train_samples_per_second": 19.117,
      "train_steps_per_second": 0.869
    }
  ],
  "max_steps": 15,
  "num_train_epochs": 15,
  "total_flos": 9.481946310220186e+16,
  "trial_name": null,
  "trial_params": null
}