trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 1135085,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 4.000405255994045e-05,
      "loss": 3.0415,
      "step": 227017
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.000801702075175e-05,
      "loss": 2.5231,
      "step": 454034
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.0011805283304773e-05,
      "loss": 2.4052,
      "step": 681051
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.001612214063264e-05,
      "loss": 2.3395,
      "step": 908068
    },
    {
      "epoch": 5.0,
      "learning_rate": 2.052709708964527e-08,
      "loss": 2.3001,
      "step": 1135085
    },
    {
      "epoch": 5.0,
      "step": 1135085,
      "total_flos": 1.9066978434981486e+18,
      "train_loss": 1.4089593235308369,
      "train_runtime": 87255.1901,
      "train_samples_per_second": 1665.125,
      "train_steps_per_second": 13.009
    }
  ],
  "max_steps": 1135085,
  "num_train_epochs": 5,
  "total_flos": 1.9066978434981486e+18,
  "trial_name": null,
  "trial_params": null
}