about adaptive_batchnorm

[yuanhaorannnnnn]

hello, I've read a lot discussions about pytorch from you, it was really helpful.
here I have a doubt about the implementation of adaptative batchnorm, why you use a * x + b * bn(x) to do it? Could you explain a bit?

[ptrblck]

The implementation should be based on Fast Image Processing with Fully-Convolutional Networks - see section 3.3 and this discussion.

Closing, but feel free to reopen in case you see any issues or post in the linked thread for more discussion.