trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 1808,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.1061946902654867,
      "grad_norm": 1.541396141052246,
      "learning_rate": 1.5575221238938054e-05,
      "loss": 0.1432,
      "step": 500
    },
    {
      "epoch": 2.2123893805309733,
      "grad_norm": 0.7666749954223633,
      "learning_rate": 1.1150442477876106e-05,
      "loss": 0.0492,
      "step": 1000
    },
    {
      "epoch": 3.3185840707964602,
      "grad_norm": 0.07702745497226715,
      "learning_rate": 6.72566371681416e-06,
      "loss": 0.0245,
      "step": 1500
    }
  ],
  "logging_steps": 500,
  "max_steps": 2260,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "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": 536680831748148.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}