Reason for scaling mask input

NVlabs / pacnet

Pixel-Adaptive Convolutional Neural Networks (CVPR '19)

https://suhangpro.github.io/pac/

Other

514 stars 79 forks source link

Reason for scaling mask input #25

Closed RSKothari closed 3 years ago

RSKothari commented 3 years ago

Hello,

I observed that you've scaled the input to the mask over here.

Is there any significance to this scaling operation?

suhangpro commented 3 years ago

This is to allow the model to train from a starting point close to the initialization. When the guidance feature is of a sufficiently small scale, its initial impact is quite small.