trainer_state.json

{
  "best_metric": 0.8650666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0661/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": 2.667667667667668e-05,
      "loss": 2.3279,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7074666666666667,
      "eval_loss": 1.2128304243087769,
      "eval_runtime": 6.8223,
      "eval_samples_per_second": 549.667,
      "eval_steps_per_second": 8.648,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 2.3353353353353354e-05,
      "loss": 0.8699,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8122666666666667,
      "eval_loss": 0.7233799695968628,
      "eval_runtime": 8.7912,
      "eval_samples_per_second": 426.563,
      "eval_steps_per_second": 6.711,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.6733522415161133,
      "learning_rate": 2.002002002002002e-05,
      "loss": 0.5338,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8336,
      "eval_loss": 0.6014339327812195,
      "eval_runtime": 6.4357,
      "eval_samples_per_second": 582.691,
      "eval_steps_per_second": 9.168,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.68538236618042,
      "learning_rate": 1.668668668668669e-05,
      "loss": 0.3598,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8517333333333333,
      "eval_loss": 0.5154938697814941,
      "eval_runtime": 5.9079,
      "eval_samples_per_second": 634.747,
      "eval_steps_per_second": 9.987,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.428194999694824,
      "learning_rate": 1.3353353353353355e-05,
      "loss": 0.2531,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8549333333333333,
      "eval_loss": 0.5047706365585327,
      "eval_runtime": 5.8741,
      "eval_samples_per_second": 638.396,
      "eval_steps_per_second": 10.044,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.691387176513672,
      "learning_rate": 1.002002002002002e-05,
      "loss": 0.1771,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8512,
      "eval_loss": 0.49734601378440857,
      "eval_runtime": 5.8958,
      "eval_samples_per_second": 636.051,
      "eval_steps_per_second": 10.007,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.5909743309020996,
      "learning_rate": 6.686686686686687e-06,
      "loss": 0.1256,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8629333333333333,
      "eval_loss": 0.48181283473968506,
      "eval_runtime": 5.808,
      "eval_samples_per_second": 645.658,
      "eval_steps_per_second": 10.158,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 4.235790729522705,
      "learning_rate": 3.353353353353354e-06,
      "loss": 0.0877,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8626666666666667,
      "eval_loss": 0.49236080050468445,
      "eval_runtime": 5.9396,
      "eval_samples_per_second": 631.351,
      "eval_steps_per_second": 9.933,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.5898157358169556,
      "learning_rate": 2.002002002002002e-08,
      "loss": 0.0662,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8650666666666667,
      "eval_loss": 0.48745450377464294,
      "eval_runtime": 6.0127,
      "eval_samples_per_second": 623.681,
      "eval_steps_per_second": 9.813,
      "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
}