trainer_state.json

{
  "best_metric": 0.9021333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0832/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": 8.730227815298347e-05,
      "loss": 2.0989,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7786666666666666,
      "eval_loss": 0.7786878943443298,
      "eval_runtime": 4.0375,
      "eval_samples_per_second": 928.79,
      "eval_steps_per_second": 14.613,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.900554656982422,
      "learning_rate": 7.95023019690139e-05,
      "loss": 0.5327,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8378666666666666,
      "eval_loss": 0.5452632308006287,
      "eval_runtime": 3.8894,
      "eval_samples_per_second": 964.15,
      "eval_steps_per_second": 15.169,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.6740777492523193,
      "learning_rate": 6.75408389678355e-05,
      "loss": 0.2453,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8696,
      "eval_loss": 0.45561861991882324,
      "eval_runtime": 3.8977,
      "eval_samples_per_second": 962.1,
      "eval_steps_per_second": 15.137,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.49221086502075195,
      "learning_rate": 5.286061811978304e-05,
      "loss": 0.1121,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8770666666666667,
      "eval_loss": 0.43212106823921204,
      "eval_runtime": 3.8925,
      "eval_samples_per_second": 963.38,
      "eval_steps_per_second": 15.157,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.4411238431930542,
      "learning_rate": 3.723229071611673e-05,
      "loss": 0.0463,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8850666666666667,
      "eval_loss": 0.42581626772880554,
      "eval_runtime": 3.9002,
      "eval_samples_per_second": 961.492,
      "eval_steps_per_second": 15.127,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.46854525804519653,
      "learning_rate": 2.2540863691266368e-05,
      "loss": 0.0138,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8874666666666666,
      "eval_loss": 0.3999105989933014,
      "eval_runtime": 4.0521,
      "eval_samples_per_second": 925.444,
      "eval_steps_per_second": 14.56,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.2064691185951233,
      "learning_rate": 1.055833996679956e-05,
      "loss": 0.0046,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8984,
      "eval_loss": 0.3850986063480377,
      "eval_runtime": 4.2226,
      "eval_samples_per_second": 888.086,
      "eval_steps_per_second": 13.973,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.028780527412891388,
      "learning_rate": 2.7299887471028354e-06,
      "loss": 0.0028,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9005333333333333,
      "eval_loss": 0.38224998116493225,
      "eval_runtime": 3.8981,
      "eval_samples_per_second": 961.999,
      "eval_steps_per_second": 15.135,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.03232070058584213,
      "learning_rate": 2.472343088255613e-11,
      "loss": 0.0025,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9021333333333333,
      "eval_loss": 0.3829493522644043,
      "eval_runtime": 5.1792,
      "eval_samples_per_second": 724.05,
      "eval_steps_per_second": 11.392,
      "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
}