Where is the XNOR operator implemented?

[yataoz]

I was looking for the implementation of XNOR operator, but couldn't find any. I only saw regular Conv layers applied to binary valued inputs and weights. Does anyone know if the convolution layer is implemented at all, which uses XNOR and bitcount to replace regular matrix multiplication?

Thanks!

[rteja1113]

I was looking for the same thing, any updates on that ?
It looks like you may need to write CUDA kernels to support that ops

[MaratZakirov]

Here I guess
https://github.com/akhauriyash/XNOR-convolution