Skip to content

hamk-uas/DifferentialConv2d

 
 

Repository files navigation

DiffConv2d (HAMK fork)

This is forked from the original code repository for the following paper:

K. Leng and J. Thiyagalingam, Padding-free Convolution based on
Preservation of Differential Characteristics of Kernels.

This repository is intended for testing and extending that work and for testing other convolutional layer border handling methods using the testing framework developed in the original work.

As a result, there is an issue in the repo for the original work, about the method being equivalent to Lagrange polynomial padding.

About

Contains the code for Borderless CNNs

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 90.1%
  • Jupyter Notebook 5.2%
  • Mathematica 4.7%