trainer_state.json

{
  "best_metric": 0.9128,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0783/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": 59.359806060791016,
      "learning_rate": 6.655434848731736e-05,
      "loss": 0.8758,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8437333333333333,
      "eval_loss": 0.516816258430481,
      "eval_runtime": 13.0947,
      "eval_samples_per_second": 286.376,
      "eval_steps_per_second": 4.506,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": Infinity,
      "learning_rate": 5.689582334051816e-05,
      "loss": 0.3046,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8674666666666667,
      "eval_loss": 0.437456339597702,
      "eval_runtime": 12.7043,
      "eval_samples_per_second": 295.175,
      "eval_steps_per_second": 4.644,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 160.52447509765625,
      "learning_rate": 4.288018105154487e-05,
      "loss": 0.1507,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8674666666666667,
      "eval_loss": 0.45091789960861206,
      "eval_runtime": 12.6127,
      "eval_samples_per_second": 297.319,
      "eval_steps_per_second": 4.678,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.802022695541382,
      "learning_rate": 2.7303772261234784e-05,
      "loss": 0.069,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8821333333333333,
      "eval_loss": 0.41845905780792236,
      "eval_runtime": 13.3568,
      "eval_samples_per_second": 280.757,
      "eval_steps_per_second": 4.417,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.005448497831821442,
      "learning_rate": 1.325169576262615e-05,
      "loss": 0.021,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9056,
      "eval_loss": 0.36593547463417053,
      "eval_runtime": 13.4388,
      "eval_samples_per_second": 279.042,
      "eval_steps_per_second": 4.39,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.12075763940811157,
      "learning_rate": 3.5071376436769933e-06,
      "loss": 0.0045,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.912,
      "eval_loss": 0.361219584941864,
      "eval_runtime": 13.2798,
      "eval_samples_per_second": 282.383,
      "eval_steps_per_second": 4.443,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.15652020275592804,
      "learning_rate": 1.2714900776311653e-10,
      "loss": 0.0025,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9128,
      "eval_loss": 0.34914448857307434,
      "eval_runtime": 13.3588,
      "eval_samples_per_second": 280.714,
      "eval_steps_per_second": 4.417,
      "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
}