search

chaofengc / PSFRGAN

PyTorch codes for "Progressive Semantic-Aware Style Transformation for Blind Face Restoration", CVPR2021
Other
370 stars 68 forks source link

what's F0 in the paper? #46

Closed tachikoma777 closed 2 years ago

tachikoma777 commented 2 years ago

There is F0 as network input, may i ask how it is used?

tachikoma777 commented 2 years ago

The paper says "It starts with a learned constant of size C1616, denoted as F0, where C is the channel size. But what's the value should be? is it the same for every image?

chaofengc commented 2 years ago

It is a learnable parameter of size C x 16 x 16.

tachikoma777 commented 2 years ago

Sorry if I bother you, but i still dont know how it is calculated, could you be more specific? Thanks a lot!