trainer_state.json

{
  "best_metric": 0.9384,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0818/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": Infinity,
      "learning_rate": 2.2545045045045048e-05,
      "loss": 0.7638,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9053333333333333,
      "eval_loss": 0.29522931575775146,
      "eval_runtime": 11.0657,
      "eval_samples_per_second": 338.886,
      "eval_steps_per_second": 5.332,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.686079502105713,
      "learning_rate": 1.5045045045045047e-05,
      "loss": 0.1164,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9248,
      "eval_loss": 0.24341832101345062,
      "eval_runtime": 10.8622,
      "eval_samples_per_second": 345.234,
      "eval_steps_per_second": 5.432,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.33166763186454773,
      "learning_rate": 7.545045045045046e-06,
      "loss": 0.0232,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9336,
      "eval_loss": 0.2262755185365677,
      "eval_runtime": 10.6138,
      "eval_samples_per_second": 353.314,
      "eval_steps_per_second": 5.559,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.02120206132531166,
      "learning_rate": 4.504504504504505e-08,
      "loss": 0.0035,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9384,
      "eval_loss": 0.2168043851852417,
      "eval_runtime": 10.9806,
      "eval_samples_per_second": 341.512,
      "eval_steps_per_second": 5.373,
      "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
}