trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 1880,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 5.32,
      "learning_rate": 3.670212765957447e-05,
      "loss": 1.4055,
      "step": 500
    },
    {
      "epoch": 10.64,
      "learning_rate": 2.340425531914894e-05,
      "loss": 0.1477,
      "step": 1000
    },
    {
      "epoch": 15.96,
      "learning_rate": 1.0106382978723404e-05,
      "loss": 0.0432,
      "step": 1500
    },
    {
      "epoch": 20.0,
      "step": 1880,
      "total_flos": 5879389501440000.0,
      "train_loss": 0.4278867336029702,
      "train_runtime": 303.7628,
      "train_samples_per_second": 197.523,
      "train_steps_per_second": 6.189
    }
  ],
  "logging_steps": 500,
  "max_steps": 1880,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 5879389501440000.0,
  "trial_name": null,
  "trial_params": null
}