search

zhanglonghao1992 / One-Shot_Free-View_Neural_Talking_Head_Synthesis

Pytorch implementation of paper "One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing"
Other
764 stars 143 forks source link

Compressing the features #12

Closed vinayak015 closed 2 years ago

vinayak015 commented 2 years ago

Before creating sparse motion (in dense_motion.py), any reason to use 1x1 conv?

zhanglonghao1992 commented 2 years ago

@vinayak015 You mean the "compress" right? Here is the explanation of the author of this paper： https://github.com/NVlabs/face-vid2vid/issues/2#issuecomment-865950115

vinayak015 commented 2 years ago

Thank you for replying. I will ask more if required. And it's a nice work :)