checkpoint-500/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.25,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_gen_len": 18.8138,
      "eval_loss": 1.9523861408233643,
      "eval_rouge1": 0.4483,
      "eval_rouge2": 0.1963,
      "eval_rougeL": 0.3162,
      "eval_rougeLsum": 0.3163,
      "eval_runtime": 2953.9269,
      "eval_samples_per_second": 0.271,
      "eval_steps_per_second": 0.135,
      "step": 400
    },
    {
      "epoch": 1.25,
      "grad_norm": 29.538663864135742,
      "learning_rate": 1.1733333333333335e-05,
      "loss": 2.2642,
      "step": 500
    }
  ],
  "logging_steps": 500,
  "max_steps": 1200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2986622080542720.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}