trainer_state.json

{
  "best_metric": 0.9157333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0577/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": 8.097160339355469,
      "learning_rate": 8.888888888888889e-05,
      "loss": 1.5227,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8184,
      "eval_loss": 0.6873378157615662,
      "eval_runtime": 8.6296,
      "eval_samples_per_second": 434.55,
      "eval_steps_per_second": 6.837,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 7.781114447781115e-05,
      "loss": 0.4538,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8552,
      "eval_loss": 0.5029574036598206,
      "eval_runtime": 7.6042,
      "eval_samples_per_second": 493.146,
      "eval_steps_per_second": 7.759,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.1253883838653564,
      "learning_rate": 6.670003336670004e-05,
      "loss": 0.2514,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8746666666666667,
      "eval_loss": 0.4564008116722107,
      "eval_runtime": 6.7504,
      "eval_samples_per_second": 555.521,
      "eval_steps_per_second": 8.74,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.749692916870117,
      "learning_rate": 5.5588922255588925e-05,
      "loss": 0.141,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8773333333333333,
      "eval_loss": 0.45086172223091125,
      "eval_runtime": 6.908,
      "eval_samples_per_second": 542.851,
      "eval_steps_per_second": 8.541,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.07380016893148422,
      "learning_rate": 4.447781114447781e-05,
      "loss": 0.0798,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8837333333333334,
      "eval_loss": 0.44294291734695435,
      "eval_runtime": 6.7271,
      "eval_samples_per_second": 557.447,
      "eval_steps_per_second": 8.771,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.1861630231142044,
      "learning_rate": 3.33667000333667e-05,
      "loss": 0.0408,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8882666666666666,
      "eval_loss": 0.43712562322616577,
      "eval_runtime": 6.7877,
      "eval_samples_per_second": 552.473,
      "eval_steps_per_second": 8.692,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.5349802374839783,
      "learning_rate": 2.225558892225559e-05,
      "loss": 0.0164,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9061333333333333,
      "eval_loss": 0.3923301696777344,
      "eval_runtime": 6.7278,
      "eval_samples_per_second": 557.389,
      "eval_steps_per_second": 8.77,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.029980842024087906,
      "learning_rate": 1.1144477811144478e-05,
      "loss": 0.0078,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9088,
      "eval_loss": 0.37738537788391113,
      "eval_runtime": 6.8158,
      "eval_samples_per_second": 550.191,
      "eval_steps_per_second": 8.656,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.09924469143152237,
      "learning_rate": 3.3366700033366706e-08,
      "loss": 0.0043,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9157333333333333,
      "eval_loss": 0.3562667667865753,
      "eval_runtime": 7.7506,
      "eval_samples_per_second": 483.834,
      "eval_steps_per_second": 7.612,
      "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
}