Hi, thanks for your great work. I am trying to use your Swin3D as a frozen feature extractor for new 3D point clouds, like how people use CLIP/DINOv2 for images.
Could you share some simple scripts or ideas about how we can achieve this with your codebase? This can be very helpful.
Hi, thanks for your great work. I am trying to use your Swin3D as a frozen feature extractor for new 3D point clouds, like how people use CLIP/DINOv2 for images.
Could you share some simple scripts or ideas about how we can achieve this with your codebase? This can be very helpful.