Sharpened Cosine Distance implementation in PyTorch

PyTorch implementation of the Sharpened Cosine Distance operator.

The core idea came from Brandon Rohrer(@brohrer) and the implementation
is based on the tf/keras implementation of Raphael Pisoni.

This implementation supports

  • 2D operation only
  • asymmetric kernels, any shape
  • CUDA / GPU

If you find this implementation useful please give it a star. Open issues for bugs/ideas.

If you are planning to build something on top of it let me know, I am always up for some good collaborations 😉

GitHub

View Github

 

 

 

To finish reading, please visit source site