trainer_state.json

{
  "best_metric": 0.8944,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0904/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.6245008707046509,
      "learning_rate": 0.0004849231551964771,
      "loss": 0.8351,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7586666666666667,
      "eval_loss": 0.8359748125076294,
      "eval_runtime": 12.5989,
      "eval_samples_per_second": 297.644,
      "eval_steps_per_second": 4.683,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0004416794553834106,
      "loss": 0.4999,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8330666666666666,
      "eval_loss": 0.569968044757843,
      "eval_runtime": 12.5266,
      "eval_samples_per_second": 299.362,
      "eval_steps_per_second": 4.71,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 16.27568817138672,
      "learning_rate": 0.00037522688315464164,
      "loss": 0.3377,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8197333333333333,
      "eval_loss": 0.6524112820625305,
      "eval_runtime": 13.0102,
      "eval_samples_per_second": 288.236,
      "eval_steps_per_second": 4.535,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.23057493567466736,
      "learning_rate": 0.0002936701006654613,
      "loss": 0.2074,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8290666666666666,
      "eval_loss": 0.6241581439971924,
      "eval_runtime": 12.6471,
      "eval_samples_per_second": 296.51,
      "eval_steps_per_second": 4.665,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.004176871385425329,
      "learning_rate": 0.00020684605953398183,
      "loss": 0.1164,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8658666666666667,
      "eval_loss": 0.5286223888397217,
      "eval_runtime": 13.0704,
      "eval_samples_per_second": 286.908,
      "eval_steps_per_second": 4.514,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.1655292958021164,
      "learning_rate": 0.00012522702050703539,
      "loss": 0.0508,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8794666666666666,
      "eval_loss": 0.5088508725166321,
      "eval_runtime": 12.6078,
      "eval_samples_per_second": 297.434,
      "eval_steps_per_second": 4.68,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.018642552196979523,
      "learning_rate": 5.8657444259997555e-05,
      "loss": 0.0163,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8917333333333334,
      "eval_loss": 0.466988205909729,
      "eval_runtime": 12.569,
      "eval_samples_per_second": 298.354,
      "eval_steps_per_second": 4.694,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.004334998317062855,
      "learning_rate": 1.5166604150571307e-05,
      "loss": 0.0051,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8941333333333333,
      "eval_loss": 0.4707319140434265,
      "eval_runtime": 12.5595,
      "eval_samples_per_second": 298.579,
      "eval_steps_per_second": 4.698,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.003827192820608616,
      "learning_rate": 1.373523937919785e-10,
      "loss": 0.0027,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8944,
      "eval_loss": 0.4748319089412689,
      "eval_runtime": 12.6916,
      "eval_samples_per_second": 295.47,
      "eval_steps_per_second": 4.649,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "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": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}