checkpoint-50/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.32,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 2.8722,
      "step": 10
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 2.7951,
      "step": 20
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 2.7847,
      "step": 30
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.0002,
      "loss": 2.7507,
      "step": 40
    },
    {
      "epoch": 0.32,
      "learning_rate": 0.0002,
      "loss": 2.7355,
      "step": 50
    }
  ],
  "max_steps": 156,
  "num_train_epochs": 1,
  "total_flos": 3.385416166146048e+16,
  "trial_name": null,
  "trial_params": null
}