trainer_state.json

{
  "best_metric": 0.9413333333333334,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0636/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": 1.2670440673828125,
      "learning_rate": 8.728616793536588e-05,
      "loss": 0.6803,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9170666666666667,
      "eval_loss": 0.28124791383743286,
      "eval_runtime": 7.6268,
      "eval_samples_per_second": 491.685,
      "eval_steps_per_second": 7.736,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.3755781948566437,
      "learning_rate": 7.947199994035401e-05,
      "loss": 0.1709,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9181333333333334,
      "eval_loss": 0.27914127707481384,
      "eval_runtime": 7.3834,
      "eval_samples_per_second": 507.893,
      "eval_steps_per_second": 7.991,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.052055858075618744,
      "learning_rate": 6.75e-05,
      "loss": 0.0811,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9256,
      "eval_loss": 0.2599741220474243,
      "eval_runtime": 7.2616,
      "eval_samples_per_second": 516.414,
      "eval_steps_per_second": 8.125,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.028287706896662712,
      "learning_rate": 5.2814167995011876e-05,
      "loss": 0.0409,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9264,
      "eval_loss": 0.2850967347621918,
      "eval_runtime": 7.5383,
      "eval_samples_per_second": 497.459,
      "eval_steps_per_second": 7.827,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.0417790524661541,
      "learning_rate": 3.7185832004988136e-05,
      "loss": 0.016,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9290666666666667,
      "eval_loss": 0.2885628342628479,
      "eval_runtime": 7.6583,
      "eval_samples_per_second": 489.664,
      "eval_steps_per_second": 7.704,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.011713231913745403,
      "learning_rate": 2.250000000000001e-05,
      "loss": 0.0062,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9368,
      "eval_loss": 0.28643545508384705,
      "eval_runtime": 7.9732,
      "eval_samples_per_second": 470.325,
      "eval_steps_per_second": 7.4,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.055833996679956e-05,
      "loss": 0.0038,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9405333333333333,
      "eval_loss": 0.2644397020339966,
      "eval_runtime": 7.5004,
      "eval_samples_per_second": 499.972,
      "eval_steps_per_second": 7.866,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.17864249646663666,
      "learning_rate": 2.7299887471028354e-06,
      "loss": 0.0027,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9389333333333333,
      "eval_loss": 0.26504039764404297,
      "eval_runtime": 7.4575,
      "eval_samples_per_second": 502.847,
      "eval_steps_per_second": 7.911,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.006555982865393162,
      "learning_rate": 2.472343088255613e-11,
      "loss": 0.0021,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9413333333333334,
      "eval_loss": 0.25548118352890015,
      "eval_runtime": 7.4454,
      "eval_samples_per_second": 503.665,
      "eval_steps_per_second": 7.924,
      "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
}