checkpoint-2264/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.09125352680370818,
  "eval_steps": 500,
  "global_step": 2264,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [],
  "logging_steps": 4962,
  "max_steps": 24810,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 2.967865494570271e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}