trainer_state.json

{
  "best_metric": 0.4808,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0430/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": Infinity,
      "learning_rate": 0.0002576576576576576,
      "loss": 3.3229,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.22613333333333333,
      "eval_loss": 2.8567075729370117,
      "eval_runtime": 7.3294,
      "eval_samples_per_second": 511.635,
      "eval_steps_per_second": 8.05,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 28.88252830505371,
      "learning_rate": 0.0002148005148005148,
      "loss": 2.7164,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.26666666666666666,
      "eval_loss": 2.6540443897247314,
      "eval_runtime": 6.4641,
      "eval_samples_per_second": 580.131,
      "eval_steps_per_second": 9.127,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 23.344541549682617,
      "learning_rate": 0.0001719433719433719,
      "loss": 2.4138,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.3592,
      "eval_loss": 2.3088865280151367,
      "eval_runtime": 7.2792,
      "eval_samples_per_second": 515.169,
      "eval_steps_per_second": 8.105,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 31.735305786132812,
      "learning_rate": 0.00012908622908622906,
      "loss": 2.1368,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.3824,
      "eval_loss": 2.2350683212280273,
      "eval_runtime": 6.3205,
      "eval_samples_per_second": 593.306,
      "eval_steps_per_second": 9.335,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.657881259918213,
      "learning_rate": 8.622908622908623e-05,
      "loss": 1.8669,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.41413333333333335,
      "eval_loss": 2.092255115509033,
      "eval_runtime": 7.3377,
      "eval_samples_per_second": 511.061,
      "eval_steps_per_second": 8.041,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 36.59032440185547,
      "learning_rate": 4.337194337194337e-05,
      "loss": 1.5741,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.464,
      "eval_loss": 1.993359923362732,
      "eval_runtime": 6.4362,
      "eval_samples_per_second": 582.645,
      "eval_steps_per_second": 9.167,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 41.84003448486328,
      "learning_rate": 5.148005148005147e-07,
      "loss": 1.1867,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.4808,
      "eval_loss": 2.035881996154785,
      "eval_runtime": 6.4132,
      "eval_samples_per_second": 584.728,
      "eval_steps_per_second": 9.2,
      "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
}