trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 7000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.71,
      "learning_rate": 4.642857142857143e-05,
      "loss": 2.4528,
      "step": 500
    },
    {
      "epoch": 1.43,
      "learning_rate": 4.2857142857142856e-05,
      "loss": 1.5218,
      "step": 1000
    },
    {
      "epoch": 2.14,
      "learning_rate": 3.928571428571429e-05,
      "loss": 1.2063,
      "step": 1500
    },
    {
      "epoch": 2.86,
      "learning_rate": 3.571428571428572e-05,
      "loss": 0.9047,
      "step": 2000
    },
    {
      "epoch": 3.57,
      "learning_rate": 3.2142857142857144e-05,
      "loss": 0.6609,
      "step": 2500
    },
    {
      "epoch": 4.29,
      "learning_rate": 2.857142857142857e-05,
      "loss": 0.5098,
      "step": 3000
    },
    {
      "epoch": 5.0,
      "learning_rate": 2.5e-05,
      "loss": 0.3947,
      "step": 3500
    },
    {
      "epoch": 5.71,
      "learning_rate": 2.1428571428571428e-05,
      "loss": 0.2421,
      "step": 4000
    },
    {
      "epoch": 6.43,
      "learning_rate": 1.785714285714286e-05,
      "loss": 0.1949,
      "step": 4500
    },
    {
      "epoch": 7.14,
      "learning_rate": 1.4285714285714285e-05,
      "loss": 0.1438,
      "step": 5000
    },
    {
      "epoch": 7.86,
      "learning_rate": 1.0714285714285714e-05,
      "loss": 0.1036,
      "step": 5500
    },
    {
      "epoch": 8.57,
      "learning_rate": 7.142857142857143e-06,
      "loss": 0.0741,
      "step": 6000
    },
    {
      "epoch": 9.29,
      "learning_rate": 3.5714285714285714e-06,
      "loss": 0.0606,
      "step": 6500
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0,
      "loss": 0.048,
      "step": 7000
    },
    {
      "epoch": 10.0,
      "step": 7000,
      "total_flos": 2.545262590643405e+16,
      "train_loss": 0.6084381016322544,
      "train_runtime": 3227.6636,
      "train_samples_per_second": 21.681,
      "train_steps_per_second": 2.169
    }
  ],
  "max_steps": 7000,
  "num_train_epochs": 10,
  "total_flos": 2.545262590643405e+16,
  "trial_name": null,
  "trial_params": null
}