wty-ustc / HairCLIP

[CVPR 2022] HairCLIP: Design Your Hair by Text and Reference Image
GNU Lesser General Public License v2.1
541 stars 68 forks source link

About modulation module #2

Closed janchen0611 closed 2 years ago

janchen0611 commented 2 years ago

Hi, Great work! But I have a question about the modulation module of mapper network. I assume the dimension of x and e should be 1x1xC. If so, what is the mean and std of x? channel-wise average? And how about the output dimensions of fr(e) and fb(e)?

Thanks.

wty-ustc commented 2 years ago

Yes. The mean and standard deviation of x are implemented by layernorm without learnable affine parameters. The output dimensions are both 1x1xC. Please follow this repository for the exact implementation, we will release the code and pre-trained model after the paper is accepted.

janchen0611 commented 2 years ago

Thanks! Looking forward to your code and model.