trainer_state.json

{
  "best_metric": 0.912,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0062/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": 3e-05,
      "loss": 0.7581,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8933333333333333,
      "eval_loss": 0.3409874439239502,
      "eval_runtime": 5.8973,
      "eval_samples_per_second": 635.886,
      "eval_steps_per_second": 10.005,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.28146952390670776,
      "learning_rate": 3e-05,
      "loss": 0.178,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9048,
      "eval_loss": 0.3019025921821594,
      "eval_runtime": 6.0594,
      "eval_samples_per_second": 618.877,
      "eval_steps_per_second": 9.737,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.759695291519165,
      "learning_rate": 3e-05,
      "loss": 0.079,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8984,
      "eval_loss": 0.331544429063797,
      "eval_runtime": 5.9796,
      "eval_samples_per_second": 627.137,
      "eval_steps_per_second": 9.867,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6622951030731201,
      "learning_rate": 3e-05,
      "loss": 0.0524,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9069333333333334,
      "eval_loss": 0.3247040808200836,
      "eval_runtime": 5.9366,
      "eval_samples_per_second": 631.671,
      "eval_steps_per_second": 9.938,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.0006717568030580878,
      "learning_rate": 3e-05,
      "loss": 0.038,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9104,
      "eval_loss": 0.3515153229236603,
      "eval_runtime": 5.8855,
      "eval_samples_per_second": 637.158,
      "eval_steps_per_second": 10.025,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 124.79778289794922,
      "learning_rate": 3e-05,
      "loss": 0.0311,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8986666666666666,
      "eval_loss": 0.4370169937610626,
      "eval_runtime": 5.9241,
      "eval_samples_per_second": 633.007,
      "eval_steps_per_second": 9.959,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.0012337745865806937,
      "learning_rate": 3e-05,
      "loss": 0.024,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.912,
      "eval_loss": 0.3805379569530487,
      "eval_runtime": 5.8419,
      "eval_samples_per_second": 641.916,
      "eval_steps_per_second": 10.099,
      "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
}