trainer_state.json

{
  "best_metric": 0.9557333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0766/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": 0.3118528723716736,
      "learning_rate": 9.698463103929542e-05,
      "loss": 0.5753,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9216,
      "eval_loss": 0.2720800042152405,
      "eval_runtime": 4.2756,
      "eval_samples_per_second": 877.061,
      "eval_steps_per_second": 13.799,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.22740942239761353,
      "learning_rate": 8.83022221559489e-05,
      "loss": 0.1199,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9365333333333333,
      "eval_loss": 0.2255595475435257,
      "eval_runtime": 4.1689,
      "eval_samples_per_second": 899.527,
      "eval_steps_per_second": 14.153,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.05884934961795807,
      "learning_rate": 7.500000000000001e-05,
      "loss": 0.0547,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9394666666666667,
      "eval_loss": 0.24168021976947784,
      "eval_runtime": 4.2529,
      "eval_samples_per_second": 881.75,
      "eval_steps_per_second": 13.873,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.024711400270462036,
      "learning_rate": 5.868240888334653e-05,
      "loss": 0.0203,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9498666666666666,
      "eval_loss": 0.20621050894260406,
      "eval_runtime": 4.3647,
      "eval_samples_per_second": 859.17,
      "eval_steps_per_second": 13.518,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 3.168926239013672,
      "learning_rate": 4.131759111665349e-05,
      "loss": 0.0098,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9485333333333333,
      "eval_loss": 0.2139752209186554,
      "eval_runtime": 4.4195,
      "eval_samples_per_second": 848.51,
      "eval_steps_per_second": 13.35,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.010251727886497974,
      "learning_rate": 2.500000000000001e-05,
      "loss": 0.0031,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9514666666666667,
      "eval_loss": 0.21923737227916718,
      "eval_runtime": 4.2617,
      "eval_samples_per_second": 879.938,
      "eval_steps_per_second": 13.844,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.1731488851999511e-05,
      "loss": 0.0017,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9538666666666666,
      "eval_loss": 0.20610813796520233,
      "eval_runtime": 4.4655,
      "eval_samples_per_second": 839.777,
      "eval_steps_per_second": 13.212,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.02142304740846157,
      "learning_rate": 3.033320830114261e-06,
      "loss": 0.0012,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9541333333333334,
      "eval_loss": 0.2050473690032959,
      "eval_runtime": 4.4627,
      "eval_samples_per_second": 840.294,
      "eval_steps_per_second": 13.221,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.018889062106609344,
      "learning_rate": 2.7470478758395702e-11,
      "loss": 0.0011,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9557333333333333,
      "eval_loss": 0.20560836791992188,
      "eval_runtime": 4.2797,
      "eval_samples_per_second": 876.236,
      "eval_steps_per_second": 13.786,
      "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
}