trainer_state.json

{
  "best_metric": 0.936,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0338/checkpoints/checkpoint-1332",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 1332,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 19.476356506347656,
      "learning_rate": 0.000225,
      "loss": 0.5574,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8864,
      "eval_loss": 0.38879671692848206,
      "eval_runtime": 4.6073,
      "eval_samples_per_second": 813.921,
      "eval_steps_per_second": 12.806,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0001502252252252252,
      "loss": 0.1836,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9154666666666667,
      "eval_loss": 0.2962331175804138,
      "eval_runtime": 4.7084,
      "eval_samples_per_second": 796.443,
      "eval_steps_per_second": 12.531,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.011229557916522026,
      "learning_rate": 7.522522522522523e-05,
      "loss": 0.0571,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9216,
      "eval_loss": 0.2846919298171997,
      "eval_runtime": 4.1147,
      "eval_samples_per_second": 911.364,
      "eval_steps_per_second": 14.339,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.01811841130256653,
      "learning_rate": 2.2522522522522522e-07,
      "loss": 0.0104,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.936,
      "eval_loss": 0.24077105522155762,
      "eval_runtime": 4.1097,
      "eval_samples_per_second": 912.483,
      "eval_steps_per_second": 14.356,
      "step": 1332
    }
  ],
  "logging_steps": 500,
  "max_steps": 1332,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 6.58965282490368e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}