trainer_state.json

{
  "best_metric": 0.9258666666666666,
  "best_model_checkpoint": "./vit_finetuned_models_dataset/CIFAR100/50_from_100/facebook_dino-vitb16/model_idx_0893/checkpoints/checkpoint-2997",
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 2997,
  "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": 4.447781114447781e-05,
      "loss": 0.7678,
      "step": 333
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8704,
      "eval_loss": 0.42061713337898254,
      "eval_runtime": 12.7045,
      "eval_samples_per_second": 295.171,
      "eval_steps_per_second": 4.644,
      "step": 333
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.684666872024536,
      "learning_rate": 3.892225558892226e-05,
      "loss": 0.2013,
      "step": 666
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8816,
      "eval_loss": 0.3966541588306427,
      "eval_runtime": 12.9166,
      "eval_samples_per_second": 290.323,
      "eval_steps_per_second": 4.568,
      "step": 666
    },
    {
      "epoch": 3.0,
      "grad_norm": 93.031494140625,
      "learning_rate": 3.33667000333667e-05,
      "loss": 0.0953,
      "step": 999
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8954666666666666,
      "eval_loss": 0.3837841749191284,
      "eval_runtime": 12.3617,
      "eval_samples_per_second": 303.356,
      "eval_steps_per_second": 4.773,
      "step": 999
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.7991158962249756,
      "learning_rate": 2.7811144477811146e-05,
      "loss": 0.0434,
      "step": 1332
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8962666666666667,
      "eval_loss": 0.3949366509914398,
      "eval_runtime": 12.5444,
      "eval_samples_per_second": 298.938,
      "eval_steps_per_second": 4.703,
      "step": 1332
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.06619105488061905,
      "learning_rate": 2.225558892225559e-05,
      "loss": 0.0205,
      "step": 1665
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9010666666666667,
      "eval_loss": 0.3665838837623596,
      "eval_runtime": 12.2862,
      "eval_samples_per_second": 305.221,
      "eval_steps_per_second": 4.802,
      "step": 1665
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.26458877325057983,
      "learning_rate": 1.6700033366700034e-05,
      "loss": 0.0069,
      "step": 1998
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9149333333333334,
      "eval_loss": 0.3754335939884186,
      "eval_runtime": 12.4282,
      "eval_samples_per_second": 301.734,
      "eval_steps_per_second": 4.747,
      "step": 1998
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.0005714398575946689,
      "learning_rate": 1.1144477811144478e-05,
      "loss": 0.0013,
      "step": 2331
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9232,
      "eval_loss": 0.3358309864997864,
      "eval_runtime": 12.3888,
      "eval_samples_per_second": 302.693,
      "eval_steps_per_second": 4.762,
      "step": 2331
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.00271156569942832,
      "learning_rate": 5.5889222555889224e-06,
      "loss": 0.0002,
      "step": 2664
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9237333333333333,
      "eval_loss": 0.33527520298957825,
      "eval_runtime": 12.6485,
      "eval_samples_per_second": 296.478,
      "eval_steps_per_second": 4.665,
      "step": 2664
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.004779680632054806,
      "learning_rate": 3.3366700033366706e-08,
      "loss": 0.0001,
      "step": 2997
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9258666666666666,
      "eval_loss": 0.3424842059612274,
      "eval_runtime": 12.6044,
      "eval_samples_per_second": 297.516,
      "eval_steps_per_second": 4.681,
      "step": 2997
    }
  ],
  "logging_steps": 500,
  "max_steps": 2997,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "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": 1.482671885603328e+19,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": null
}