trainer_state.json

{
  "best_metric": 0.8328,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_vit-mae-base/model_idx_0303/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": 0.00026676676676676677,
      "loss": 2.9889,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3834666666666667,
      "eval_loss": 2.2727224826812744,
      "eval_runtime": 12.2571,
      "eval_samples_per_second": 305.945,
      "eval_steps_per_second": 4.814,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 0.0002335335335335335,
      "loss": 1.6016,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6573333333333333,
      "eval_loss": 1.1764380931854248,
      "eval_runtime": 10.9755,
      "eval_samples_per_second": 341.671,
      "eval_steps_per_second": 5.376,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 30.939237594604492,
      "learning_rate": 0.00020020020020020016,
      "loss": 0.8415,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7285333333333334,
      "eval_loss": 0.891646683216095,
      "eval_runtime": 11.0464,
      "eval_samples_per_second": 339.479,
      "eval_steps_per_second": 5.341,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 9.534031867980957,
      "learning_rate": 0.00016686686686686686,
      "loss": 0.475,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7709333333333334,
      "eval_loss": 0.7716444134712219,
      "eval_runtime": 11.355,
      "eval_samples_per_second": 330.251,
      "eval_steps_per_second": 5.196,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 10.020750999450684,
      "learning_rate": 0.00013353353353353354,
      "loss": 0.2536,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7872,
      "eval_loss": 0.7351541519165039,
      "eval_runtime": 11.1394,
      "eval_samples_per_second": 336.643,
      "eval_steps_per_second": 5.297,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 3.4989047050476074,
      "learning_rate": 0.0001002002002002002,
      "loss": 0.116,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8013333333333333,
      "eval_loss": 0.7216194272041321,
      "eval_runtime": 10.9681,
      "eval_samples_per_second": 341.9,
      "eval_steps_per_second": 5.379,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.045611247420310974,
      "learning_rate": 6.686686686686686e-05,
      "loss": 0.0471,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8178666666666666,
      "eval_loss": 0.7181538939476013,
      "eval_runtime": 10.9334,
      "eval_samples_per_second": 342.985,
      "eval_steps_per_second": 5.396,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 6.0000834465026855,
      "learning_rate": 3.3533533533533535e-05,
      "loss": 0.0168,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.8274666666666667,
      "eval_loss": 0.6775506138801575,
      "eval_runtime": 11.2482,
      "eval_samples_per_second": 333.386,
      "eval_steps_per_second": 5.245,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.027188889682292938,
      "learning_rate": 2.002002002002002e-07,
      "loss": 0.0065,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8328,
      "eval_loss": 0.6713429689407349,
      "eval_runtime": 11.2613,
      "eval_samples_per_second": 332.999,
      "eval_steps_per_second": 5.239,
      "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
}