Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
FFT backend for Filter2d
See original GitHub issue🚀 Feature
An FFT backend for kornia.filters.filter2d.
Motivation
kornia.filters.filter2d is very slow for large kernels as it currently only performs convolutions in the spatial domain with conv2d.
Pitch
Convolution in the spatial domain is equivalent to multiplication in the Fourier domain. For large kernels, transforming both the kernel and the image to the fourier domain, multiplying, and inverting the FFT should yield a much faster filter2d implementation for large kernels (in my experience - filters larger than 10x10 pixels probably benefit from a fourier transform).
Alternatives
Additional context
Issue Analytics
- State:
- Created a year ago
- Comments:5 (3 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
@oliland that would be an awesome contribution. I guess we are targeting something like this: https://stackoverflow.com/questions/60561933/verify-convolution-theorem-using-pytorch
A benchmark should be included. I also noticed some slow downs when using large kenels in gaussian blur and similar
Yes, but for others it would be even more complicated, we would have to have many ifs per backend per kernels size then I think that CPU conv is done via im2col https://discuss.pytorch.org/t/how-was-conv2d-implemented-in-pytorch/35223