trainer_state.json

{
  "best_metric": 0.9362666666666667,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0496/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": 43.38261413574219,
      "learning_rate": 7.722007722007723e-05,
      "loss": 0.8318,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8664,
      "eval_loss": 0.44653555750846863,
      "eval_runtime": 8.667,
      "eval_samples_per_second": 432.675,
      "eval_steps_per_second": 6.807,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.464056968688965,
      "learning_rate": 6.436293436293436e-05,
      "loss": 0.2873,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8693333333333333,
      "eval_loss": 0.442794531583786,
      "eval_runtime": 7.8618,
      "eval_samples_per_second": 476.99,
      "eval_steps_per_second": 7.505,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.3221495151519775,
      "learning_rate": 5.1505791505791514e-05,
      "loss": 0.1477,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8845333333333333,
      "eval_loss": 0.42923352122306824,
      "eval_runtime": 7.6796,
      "eval_samples_per_second": 488.306,
      "eval_steps_per_second": 7.683,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.028060806915163994,
      "learning_rate": 3.864864864864865e-05,
      "loss": 0.0734,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9061333333333333,
      "eval_loss": 0.36935222148895264,
      "eval_runtime": 8.0612,
      "eval_samples_per_second": 465.194,
      "eval_steps_per_second": 7.319,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.5288587808609009,
      "learning_rate": 2.5791505791505793e-05,
      "loss": 0.0324,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9218666666666666,
      "eval_loss": 0.3150714933872223,
      "eval_runtime": 7.8374,
      "eval_samples_per_second": 478.472,
      "eval_steps_per_second": 7.528,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.04076455533504486,
      "learning_rate": 1.2934362934362936e-05,
      "loss": 0.0107,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9338666666666666,
      "eval_loss": 0.2726023495197296,
      "eval_runtime": 8.224,
      "eval_samples_per_second": 455.984,
      "eval_steps_per_second": 7.174,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 7.075593948364258,
      "learning_rate": 7.722007722007723e-08,
      "loss": 0.0017,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9362666666666667,
      "eval_loss": 0.25914040207862854,
      "eval_runtime": 8.1013,
      "eval_samples_per_second": 462.888,
      "eval_steps_per_second": 7.283,
      "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
}