trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.43,
      "learning_rate": 4.2857142857142856e-05,
      "loss": 6.5128,
      "step": 500
    },
    {
      "epoch": 2.86,
      "learning_rate": 3.571428571428572e-05,
      "loss": 3.5373,
      "step": 1000
    },
    {
      "epoch": 4.29,
      "learning_rate": 2.857142857142857e-05,
      "loss": 1.21,
      "step": 1500
    },
    {
      "epoch": 5.71,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 0.9726,
      "step": 2000
    },
    {
      "epoch": 7.14,
      "learning_rate": 1.4285714285714285e-05,
      "loss": 0.4473,
      "step": 2500
    },
    {
      "epoch": 8.57,
      "learning_rate": 7.142857142857143e-06,
      "loss": 0.2415,
      "step": 3000
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0,
      "loss": 0.2206,
      "step": 3500
    },
    {
      "epoch": 10.0,
      "step": 3500,
      "total_flos": 1.389991530676224e+16,
      "train_loss": 1.8774449462890626,
      "train_runtime": 1626.4344,
      "train_samples_per_second": 21.519,
      "train_steps_per_second": 2.152
    }
  ],
  "max_steps": 3500,
  "num_train_epochs": 10,
  "total_flos": 1.389991530676224e+16,
  "trial_name": null,
  "trial_params": null
}