trainer_state.json

{
  "best_metric": 0.9341333333333334,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/google_vit-base-patch16-224/model_idx_0795/checkpoints/checkpoint-2331",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2331,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.2886744737625122,
      "learning_rate": 0.0002571428571428571,
      "loss": 0.5754,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8736,
      "eval_loss": 0.42436009645462036,
      "eval_runtime": 10.4078,
      "eval_samples_per_second": 360.308,
      "eval_steps_per_second": 5.669,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.09006724506616592,
      "learning_rate": 0.00021428571428571427,
      "loss": 0.2405,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9,
      "eval_loss": 0.34577903151512146,
      "eval_runtime": 10.0052,
      "eval_samples_per_second": 374.806,
      "eval_steps_per_second": 5.897,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.9565986394882202,
      "learning_rate": 0.0001714285714285714,
      "loss": 0.1308,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9037333333333334,
      "eval_loss": 0.3441358804702759,
      "eval_runtime": 10.0361,
      "eval_samples_per_second": 373.65,
      "eval_steps_per_second": 5.879,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.007625298108905554,
      "learning_rate": 0.00012857142857142855,
      "loss": 0.0615,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9210666666666667,
      "eval_loss": 0.29041439294815063,
      "eval_runtime": 10.888,
      "eval_samples_per_second": 344.415,
      "eval_steps_per_second": 5.419,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.826369762420654,
      "learning_rate": 8.57142857142857e-05,
      "loss": 0.0249,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.928,
      "eval_loss": 0.29922983050346375,
      "eval_runtime": 9.9878,
      "eval_samples_per_second": 375.459,
      "eval_steps_per_second": 5.907,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.021261410787701607,
      "learning_rate": 4.285714285714285e-05,
      "loss": 0.0067,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9317333333333333,
      "eval_loss": 0.2704765796661377,
      "eval_runtime": 10.0357,
      "eval_samples_per_second": 373.666,
      "eval_steps_per_second": 5.879,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": NaN,
      "learning_rate": 1.2870012870012868e-07,
      "loss": 0.0024,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9341333333333334,
      "eval_loss": 0.2674747407436371,
      "eval_runtime": 10.4477,
      "eval_samples_per_second": 358.929,
      "eval_steps_per_second": 5.647,
      "step": 2331
    }
  ],
  "logging_steps": 500,
  "max_steps": 2331,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "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": 1.153189244358144e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}