search

deeptibhegde / CLIP-goes-3D

Official code release of "CLIP goes 3D: Leveraging Prompt Tuning for Language Grounded 3D Recognition"
https://jeya-maria-jose.github.io/cg3d-web/
218 stars 13 forks source link

color information #19

Closed yushiangw closed 10 months ago

yushiangw commented 10 months ago

Have a general question. Do you try to use color information when you train the 3D point cloud encoder? Thanks!

deeptibhegde commented 10 months ago

No, the color information is not used to train the 3D encoder.