trainer_state.json

{
  "best_metric": 0.9213333333333333,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0780/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": NaN,
      "learning_rate": 6.657472928225117e-05,
      "loss": 0.7215,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8573333333333333,
      "eval_loss": 0.4958763122558594,
      "eval_runtime": 10.7869,
      "eval_samples_per_second": 347.643,
      "eval_steps_per_second": 5.47,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.1923331767320633,
      "learning_rate": 5.689582334051816e-05,
      "loss": 0.2378,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8669333333333333,
      "eval_loss": 0.48273804783821106,
      "eval_runtime": 10.413,
      "eval_samples_per_second": 360.128,
      "eval_steps_per_second": 5.666,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.113832950592041,
      "learning_rate": 4.288018105154487e-05,
      "loss": 0.1092,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8837333333333334,
      "eval_loss": 0.42864200472831726,
      "eval_runtime": 10.2465,
      "eval_samples_per_second": 365.98,
      "eval_steps_per_second": 5.758,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.8119053840637207,
      "learning_rate": 2.7303772261234784e-05,
      "loss": 0.0377,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.896,
      "eval_loss": 0.4303550720214844,
      "eval_runtime": 10.4123,
      "eval_samples_per_second": 360.152,
      "eval_steps_per_second": 5.666,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.0022301641292870045,
      "learning_rate": 1.325169576262615e-05,
      "loss": 0.0107,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.908,
      "eval_loss": 0.3720003366470337,
      "eval_runtime": 10.6071,
      "eval_samples_per_second": 353.537,
      "eval_steps_per_second": 5.562,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.6125485823722556e-05,
      "learning_rate": 3.5071376436769933e-06,
      "loss": 0.0021,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9178666666666667,
      "eval_loss": 0.3436310589313507,
      "eval_runtime": 10.5066,
      "eval_samples_per_second": 356.918,
      "eval_steps_per_second": 5.616,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.0002159721771022305,
      "learning_rate": 1.2714900776311653e-10,
      "loss": 0.0008,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9213333333333333,
      "eval_loss": 0.3362419307231903,
      "eval_runtime": 10.3841,
      "eval_samples_per_second": 361.13,
      "eval_steps_per_second": 5.682,
      "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
}