search

zyhbili / Pose2Img

A warping based image translation model focusing on upper body synthesis.
35 stars 5 forks source link

pretrained weights of Luo and Xing? #5

Open zhanganny opened 5 months ago

zhanganny commented 5 months ago

Sorry to bother you. We want to evaluate the video generation effect of Speech Drives Templates. Could you please provide the training weights of Luo Xiang and Song Xing

zyhbili commented 5 months ago

Sorry, i lost the checkpoints since i have graduated from school. SDT does not focus on the video generation, and the diffusion-based models (like animate-anyone), seem to completely outperform the warping-based method in this repo. If you want to visualize the skeleton produced by SDT, maybe you could try "通义千问" or champ. If you need more training details, please contact my email 1372241206@qq.com.