yohanshin / WHAM

MIT License
565 stars 58 forks source link

About the provided pretrained models #72

Open ColinZhou1993 opened 3 months ago

ColinZhou1993 commented 3 months ago

Great work! Want to know if you could release the pretrained models with different backbones( WHAM (Res)/(HR)) in the paper. I want to run the demo in my own pc in real-time(>30fps) way, but the WHAM(ViT) is time costly. Thanks a lot!

yohanshin commented 2 months ago

Hi Colin, yes I will release the other two backbones as well. But as you can see in our supplementary materials, backbone is not the bottleneck in WHAM. You can consider using lightweighted keypoints detection model instead of ViTPose. The performance will slightly degenerate btw.