JeremyCJM / DiffSHEG

[CVPR'24] DiffSHEG: A Diffusion-Based Approach for Real-Time Speech-driven Holistic 3D Expression and Gesture Generation
https://jeremycjm.github.io/proj/DiffSHEG/
BSD 3-Clause "New" or "Revised" License
112 stars 9 forks source link

dataset #20

Open user99089 opened 1 month ago

user99089 commented 1 month ago

Hello, I am glad that you can share your great work. As a beginner, I have a request: Could you provide a Google Drive link to the BEAT dataset you used, and indicate where the dataset should be placed? The main reason is that I couldn't find some of the facial weights (.bin) you used in the BEAT dataset and other works are usually like this.

Thank you very much!