trainer_state.json

{
  "best_metric": 0.9082666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0222/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": Infinity,
      "learning_rate": 8.895562228895563e-05,
      "loss": 1.5132,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7914666666666667,
      "eval_loss": 0.7295359969139099,
      "eval_runtime": 10.6706,
      "eval_samples_per_second": 351.434,
      "eval_steps_per_second": 5.529,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.704954624176025,
      "learning_rate": 7.784451117784451e-05,
      "loss": 0.4842,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.848,
      "eval_loss": 0.5258737802505493,
      "eval_runtime": 12.4725,
      "eval_samples_per_second": 300.662,
      "eval_steps_per_second": 4.73,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.1408126801252365,
      "learning_rate": 6.67334000667334e-05,
      "loss": 0.2702,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8616,
      "eval_loss": 0.47964218258857727,
      "eval_runtime": 10.6247,
      "eval_samples_per_second": 352.953,
      "eval_steps_per_second": 5.553,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.3452495038509369,
      "learning_rate": 5.562228895562229e-05,
      "loss": 0.1581,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8752,
      "eval_loss": 0.42705994844436646,
      "eval_runtime": 11.4073,
      "eval_samples_per_second": 328.738,
      "eval_steps_per_second": 5.172,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.22586296498775482,
      "learning_rate": 4.451117784451118e-05,
      "loss": 0.0895,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8792,
      "eval_loss": 0.4485592842102051,
      "eval_runtime": 10.516,
      "eval_samples_per_second": 356.599,
      "eval_steps_per_second": 5.61,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.3898897171020508,
      "learning_rate": 3.340006673340007e-05,
      "loss": 0.0442,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8890666666666667,
      "eval_loss": 0.4157477617263794,
      "eval_runtime": 10.3923,
      "eval_samples_per_second": 360.844,
      "eval_steps_per_second": 5.677,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.3365578055381775,
      "learning_rate": 2.2288955622288957e-05,
      "loss": 0.0218,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8936,
      "eval_loss": 0.42556512355804443,
      "eval_runtime": 10.8702,
      "eval_samples_per_second": 344.981,
      "eval_steps_per_second": 5.428,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.02917780540883541,
      "learning_rate": 1.1177844511177845e-05,
      "loss": 0.0091,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9026666666666666,
      "eval_loss": 0.3997747600078583,
      "eval_runtime": 11.8492,
      "eval_samples_per_second": 316.478,
      "eval_steps_per_second": 4.979,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.1620771586894989,
      "learning_rate": 6.673340006673341e-08,
      "loss": 0.004,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9082666666666667,
      "eval_loss": 0.39737191796302795,
      "eval_runtime": 10.6271,
      "eval_samples_per_second": 352.872,
      "eval_steps_per_second": 5.552,
      "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
}