trainer_state.json

{
  "best_metric": 0.9461333333333334,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0282/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": 0.9355396032333374,
      "learning_rate": 8.888888888888889e-05,
      "loss": 0.6222,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9178666666666667,
      "eval_loss": 0.28708913922309875,
      "eval_runtime": 11.2288,
      "eval_samples_per_second": 333.963,
      "eval_steps_per_second": 5.254,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.07045841962099075,
      "learning_rate": 7.777777777777778e-05,
      "loss": 0.1467,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9237333333333333,
      "eval_loss": 0.24740412831306458,
      "eval_runtime": 11.3975,
      "eval_samples_per_second": 329.019,
      "eval_steps_per_second": 5.177,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.3695173263549805,
      "learning_rate": 6.666666666666667e-05,
      "loss": 0.0747,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.928,
      "eval_loss": 0.24054840207099915,
      "eval_runtime": 11.3945,
      "eval_samples_per_second": 329.105,
      "eval_steps_per_second": 5.178,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.01600927859544754,
      "learning_rate": 5.555555555555556e-05,
      "loss": 0.0411,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9314666666666667,
      "eval_loss": 0.25431570410728455,
      "eval_runtime": 11.3524,
      "eval_samples_per_second": 330.326,
      "eval_steps_per_second": 5.197,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.4718668758869171,
      "learning_rate": 4.4444444444444447e-05,
      "loss": 0.021,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9362666666666667,
      "eval_loss": 0.25286614894866943,
      "eval_runtime": 11.1609,
      "eval_samples_per_second": 335.994,
      "eval_steps_per_second": 5.286,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.008586987853050232,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.013,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9386666666666666,
      "eval_loss": 0.26306167244911194,
      "eval_runtime": 11.3498,
      "eval_samples_per_second": 330.403,
      "eval_steps_per_second": 5.198,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 2.225558892225559e-05,
      "loss": 0.0055,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9408,
      "eval_loss": 0.26461565494537354,
      "eval_runtime": 12.2182,
      "eval_samples_per_second": 306.92,
      "eval_steps_per_second": 4.829,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.006249438039958477,
      "learning_rate": 1.1144477811144478e-05,
      "loss": 0.0024,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9424,
      "eval_loss": 0.2521492540836334,
      "eval_runtime": 11.0255,
      "eval_samples_per_second": 340.121,
      "eval_steps_per_second": 5.351,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.024063115939497948,
      "learning_rate": 3.3366700033366706e-08,
      "loss": 0.0014,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9461333333333334,
      "eval_loss": 0.2487383782863617,
      "eval_runtime": 11.0103,
      "eval_samples_per_second": 340.59,
      "eval_steps_per_second": 5.359,
      "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
}