search

LinZhekai / X-Oscar

About Official repository for "X-Oscar: A Progressive Framework for High-quality Text-guided 3D Animatable Avatar Generation"
46 stars 2 forks source link

How to use AIST/talk show/motion diffusion to animation? #6

Open yejr0229 opened 2 weeks ago

yejr0229 commented 2 weeks ago

Hi, could you please tell me the specific steps about use these data to animation, I'm confused about how to manage the dataset

LinZhekai commented 2 weeks ago

After preparing the required dataset, we implement animation through the apps/anime.py file. For talkshow and mdm, you can pass in the file address through --talkshow_file or --mdm_file. For the AIST dataset, you need to change the aist_dir in the anime.py file and select different actions by changing the motion_name. In the forward_aist function, the camera data and action data from the AIST dataset will be loaded to complete the animation of the 3D avatar and save it as a video.