Understanding weight-magnitude hyperparameters in training binary networks

This is the PyTorch implementation of accepted ICLR 2023 paper: Understanding weight-magnitude hyperparameters in training binary networks

Citation

  @inproceedings{
  quist2023understanding,
  title={Understanding weight-magnitude hyperparameters in training  binary networks},
  author={Joris Quist and Yunqiang Li and Jan van Gemert},
  booktitle={International Conference on Learning Representations},
  year={2023},
  url={https://openreview.net/forum?id=uBKBoix9NXa}
  }