zhou-linpeng

> Without data preprocessing, a random picture is used as ref_image and the provided motion_6 for inference. The result is as follow. The consistency of the character's movements is very...

> > > Without data preprocessing, a random picture is used as ref_image and the provided motion_6 for inference. The result is as follow. The consistency of the character's movements...

A preliminary Champ was implemented, including aligning the 3D pose with the input image. The entire forward process took about 6 minutes. https://c933b786ceedee6c17.gradio.live https://github.com/fudan-generative-vision/champ/assets/147801386/a53d5a16-b304-4a3e-bef6-0d544ee53b33

you can use deepspeed to reduce gpu memory

where is the 4D-humans git, I cant find it, can you share it with us, thanks! > inference_smpl.py is located at "/4D-Humans" folder which you need to git clone separately...

> Hi @zhou-linpeng sorry that this repo has been deleted cause it's not convenient for users to set up the environment. You can git clone https://github.com/Leoooo333/4D-Humans, they are the same....

> Hi @zhou-linpeng sorry that this repo has been deleted cause it's not convenient for users to set up the environment. You can git clone https://github.com/Leoooo333/4D-Humans, they are the same....