trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.983461962513782,
  "global_step": 1130,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.9982300884955754e-05,
      "loss": 2.3012,
      "step": 1
    },
    {
      "epoch": 2.21,
      "learning_rate": 1.1150442477876106e-05,
      "loss": 0.9544,
      "step": 500
    },
    {
      "epoch": 4.41,
      "learning_rate": 2.3008849557522127e-06,
      "loss": 0.673,
      "step": 1000
    },
    {
      "epoch": 4.98,
      "step": 1130,
      "total_flos": 6.709821425949082e+16,
      "train_loss": 0.7888860991570802,
      "train_runtime": 4654.6916,
      "train_samples_per_second": 1.947,
      "train_steps_per_second": 0.243
    }
  ],
  "max_steps": 1130,
  "num_train_epochs": 5,
  "total_flos": 6.709821425949082e+16,
  "trial_name": null,
  "trial_params": null
}