/Forgetting-to-Witness

Primary LanguagePython

Forgetting to Witness: Efficient Federated Unlearning and Its Visible Evaluation

PyTorch code for the paper "Forgetting to Witness: Efficient Federated Unlearning and Its Visible Evaluation".

1 Requirements

We recommended the following dependencies.

  • python==3.8
  • pytorch==1.11.0
  • torchvision==0.12.0
  • numpy==1.22.4

For more recommended dependencies, please refer to the file requirements.txt.

2 How to use

2.1 Training new models

We use the CIFAR-10 dataset as an example.

Run train_cifar10.py to obtain the trained models:

python train_cifar10.py

You can specify the number of training epochs by setting the epoch_num in the file. If necessary, please modify the path to the dataset; otherwise, the code will automatically download the dataset to the default path when executed.

2.2 Unlearning

Run unlearning_cifar10.py to perform the unlearning process:

python unlearning_cifar10.py

To customize the unlearning process, please make the following adjustments in the file:

  • Modify the epochs value to specify the number of unlearning epochs.
  • Modify the your_backdoored_model_path to set the path for the backdoored model.
  • Modify the forget_Proportion to define the size of the forgetting dataset.

These configurations will allow you to tailor the unlearning process to your specific requirements. Ensure that the values are correctly set before running the training script to avoid any errors or unexpected results.

2.3 Visible Evaluation

We use the MNIST dataset as an example.

Run GAN-MNIST.py to perform the visible evaluation process:

python GAN-MNIST.py

To customize the evaluation process, please make the following adjustments in the file:

  • Modify the num_epochs value to specify the number of unlearning epochs.
  • Modify the classifier_path to the trained classifier checkpoint.
  • Modify the generator_save_path to the save path for generator checkpoint.

These configurations will allow you to tailor the evaluation process to your specific requirements. Ensure that the values are correctly set before running the training script to avoid any errors or unexpected results.

3 Code Reference

For detailed code explanations and best practices, please refer to