trainer_state.json

{
  "best_metric": 0.9218666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0813/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.893980979919434,
      "learning_rate": 8.559608050575152e-05,
      "loss": 0.8893,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8112,
      "eval_loss": 0.6233060956001282,
      "eval_runtime": 11.1624,
      "eval_samples_per_second": 335.949,
      "eval_steps_per_second": 5.286,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.2220192402601242,
      "learning_rate": 7.31517728663805e-05,
      "loss": 0.3382,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8725333333333334,
      "eval_loss": 0.43561968207359314,
      "eval_runtime": 11.1443,
      "eval_samples_per_second": 336.495,
      "eval_steps_per_second": 5.294,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 15.321968078613281,
      "learning_rate": 5.5131661351986275e-05,
      "loss": 0.1754,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8834666666666666,
      "eval_loss": 0.42498284578323364,
      "eval_runtime": 11.3541,
      "eval_samples_per_second": 330.278,
      "eval_steps_per_second": 5.196,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 37.67421340942383,
      "learning_rate": 3.5104850050159014e-05,
      "loss": 0.0878,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8976,
      "eval_loss": 0.3912941515445709,
      "eval_runtime": 10.924,
      "eval_samples_per_second": 343.282,
      "eval_steps_per_second": 5.401,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8125289082527161,
      "learning_rate": 1.703789455194791e-05,
      "loss": 0.0385,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9098666666666667,
      "eval_loss": 0.34108951687812805,
      "eval_runtime": 11.2737,
      "eval_samples_per_second": 332.634,
      "eval_steps_per_second": 5.233,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.000431274704169482,
      "learning_rate": 4.509176970441849e-06,
      "loss": 0.0112,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9141333333333334,
      "eval_loss": 0.32491084933280945,
      "eval_runtime": 11.5109,
      "eval_samples_per_second": 325.778,
      "eval_steps_per_second": 5.126,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.009147590026259422,
      "learning_rate": 1.6347729569543557e-10,
      "loss": 0.003,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9218666666666666,
      "eval_loss": 0.32436272501945496,
      "eval_runtime": 11.1152,
      "eval_samples_per_second": 337.375,
      "eval_steps_per_second": 5.308,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}