huawei-noah / AdderNet

Code for paper " AdderNet: Do We Really Need Multiplications in Deep Learning?"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

size of alpha in AdderSR

syncerpn opened this issue · comments

I have read the AdderSR paper. It's very interesting.
I have one question about the alpha parameter.
Assuming the output feature maps is a w x h x c tensor, what is the shape of alpha? Is is a single param or a vector of c elements?

Thank you!

It is a single param.