trainer_state.json

{
  "best_metric": 0.9312,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0983/checkpoints/checkpoint-1332",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 1332,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 4.689077854156494,
      "learning_rate": 7.689476527896773e-05,
      "loss": 0.8136,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8592,
      "eval_loss": 0.4752389192581177,
      "eval_runtime": 13.3853,
      "eval_samples_per_second": 280.158,
      "eval_steps_per_second": 4.408,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.0890088602900505,
      "learning_rate": 4.510613478854492e-05,
      "loss": 0.2551,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.888,
      "eval_loss": 0.36733919382095337,
      "eval_runtime": 13.3677,
      "eval_samples_per_second": 280.528,
      "eval_steps_per_second": 4.414,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 25.74648666381836,
      "learning_rate": 1.3255331978432106e-05,
      "loss": 0.0756,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9194666666666667,
      "eval_loss": 0.2700423300266266,
      "eval_runtime": 13.3328,
      "eval_samples_per_second": 281.262,
      "eval_steps_per_second": 4.425,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": Infinity,
      "learning_rate": 5.006485928676918e-10,
      "loss": 0.0147,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9312,
      "eval_loss": 0.2366323173046112,
      "eval_runtime": 13.9187,
      "eval_samples_per_second": 269.423,
      "eval_steps_per_second": 4.239,
      "step": 1332
    }
  ],
  "logging_steps": 500,
  "max_steps": 1332,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 6.58965282490368e+18,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}