Skip to content

ahmdtaha/svmax

Repository files navigation

SVMax: A Feature Embedding Regularizer

The official implementation of SVMax: A Feature Embedding Regularizer

Requirements

  • Python 3+ [Tested on 3.7]
  • Tensorflow 2.X [Tested on 2.0]

Usage example

The toy mnist script mnist_embedding.py is straight-forward to run -- no external dependencies. Set the variable svmax_enabled to turn SVMax On/Off.

Release History

  • 1.0.0
    • Add a toy Tensorflow example on Feb 20, 2021
    • Add a toy PyTorch example on Feb 21, 2021
    • Add a toy MNIST project to the 2D unit circle on Mar 5, 2021

Contributor list

  1. Ahmed Taha
  2. Alex Hanson

Citation

@article{taha2021svmax,
  title={SVMax: A Feature Embedding Regularizer},
  author={Taha, Ahmed and Hanson, Alex and Shrivastava, Abhinav and Davis, Larry},
  journal={arXiv preprint arXiv:2103.02770},
  year={2021}
}