trainer_state.json

{
  "best_metric": 0.912,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0117/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": Infinity,
      "learning_rate": 4.446112779446113e-05,
      "loss": 1.947,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7986666666666666,
      "eval_loss": 0.8279477953910828,
      "eval_runtime": 13.1675,
      "eval_samples_per_second": 284.791,
      "eval_steps_per_second": 4.481,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 3.892225558892226e-05,
      "loss": 0.5914,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8493333333333334,
      "eval_loss": 0.5338485836982727,
      "eval_runtime": 12.7342,
      "eval_samples_per_second": 294.482,
      "eval_steps_per_second": 4.633,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 101.2293701171875,
      "learning_rate": 3.33667000333667e-05,
      "loss": 0.3631,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8597333333333333,
      "eval_loss": 0.4604315757751465,
      "eval_runtime": 13.4126,
      "eval_samples_per_second": 279.587,
      "eval_steps_per_second": 4.399,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 25.476119995117188,
      "learning_rate": 2.7811144477811146e-05,
      "loss": 0.2397,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8762666666666666,
      "eval_loss": 0.42145416140556335,
      "eval_runtime": 13.2671,
      "eval_samples_per_second": 282.654,
      "eval_steps_per_second": 4.447,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.607358992099762,
      "learning_rate": 2.225558892225559e-05,
      "loss": 0.1587,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8917333333333334,
      "eval_loss": 0.3632453680038452,
      "eval_runtime": 13.3151,
      "eval_samples_per_second": 281.636,
      "eval_steps_per_second": 4.431,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.5837167501449585,
      "learning_rate": 1.6700033366700034e-05,
      "loss": 0.1057,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8970666666666667,
      "eval_loss": 0.3490012288093567,
      "eval_runtime": 13.2182,
      "eval_samples_per_second": 283.699,
      "eval_steps_per_second": 4.464,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.09330987930297852,
      "learning_rate": 1.1144477811144478e-05,
      "loss": 0.0646,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8936,
      "eval_loss": 0.36935490369796753,
      "eval_runtime": 13.2372,
      "eval_samples_per_second": 283.294,
      "eval_steps_per_second": 4.457,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.5630838871002197,
      "learning_rate": 5.5889222555889224e-06,
      "loss": 0.042,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.904,
      "eval_loss": 0.35329005122184753,
      "eval_runtime": 13.4995,
      "eval_samples_per_second": 277.789,
      "eval_steps_per_second": 4.371,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 3.9800925254821777,
      "learning_rate": 3.3366700033366706e-08,
      "loss": 0.0248,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.912,
      "eval_loss": 0.3244766294956207,
      "eval_runtime": 13.3776,
      "eval_samples_per_second": 280.32,
      "eval_steps_per_second": 4.41,
      "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
}