search

hongfz16 / AvatarCLIP

[SIGGRAPH 2022 Journal Track] AvatarCLIP: Zero-Shot Text-Driven Generation and Animation of 3D Avatars
https://hongfz16.github.io/projects/AvatarCLIP.html
Other
1.06k stars 93 forks source link

Motion Generation #2

Open sard0r opened 2 years ago

sard0r commented 2 years ago

Congratulations @hongfz16 very impactful work to the community. I was amazed by your work. However, I wanted to ask if you can upload Motion generation code and give a guideline of how to use it. Thank you

mingyuan-zhang commented 2 years ago

Hi @sard0r , thanks for your appreciation of our work. I'm tidying up the code and writing the guidelines for the Motion Generation part. This part will be uploaded in two weeks. I'm sorry for the late update.

Jameshskelton commented 2 years ago

Yes, I am writing a tutorial to run the code on remote GPU vms, and I think it would be much better with the motion generation portion. @mingyuan-zhang, do you think it will be ready soon? Thank you for sharing your work.

ai1urus commented 2 years ago

@mingyuan-zhang Thank you for your amazing work. It's been three weeks. Will you upload the motion generation code rencently?

mingyuan-zhang commented 2 years ago

We have released Motion Generation Part now. There are some new installation steps here.

  1. requirements.txt​ has been updated. You may pip install -r requirements​ again.
  2. run conda install -c menpo osmesa​ to support offscreen rendering
  3. download data from https://drive.google.com/drive/folders/1TSyeT8MwH5EVQRbNGRVkWsA4Y9Y6dRbk?usp=sharing and VPoser v2.0 pretrained weight from https://smpl-x.is.tue.mpg.de/download.php

You may install the new libraries and download new data. Then you can run the code of motion generation part. Feel free to ask me if you have more questions.