Pointcept / Pointcept

Pointcept: a codebase for point cloud perception research. Latest works: PTv3 (CVPR'24 Oral), PPT (CVPR'24), OA-CNNs (CVPR'24), MSC (CVPR'23)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How can I run my dataset?

CacatuaAlan opened this issue · comments

Snipaste_2024-05-29_17-23-41
Hello! I have a dataset like the image. I want to know how to run in ptv3. Thanks a lot!

Hi, you can refer to one of our processed data and arrange your data with a similar structure.
https://huggingface.co/datasets/Pointcept/scannet-compressed

Hi, you can refer to one of our processed data and arrange your data with a similar structure. https://huggingface.co/datasets/Pointcept/scannet-compressed

It works! Thanks! By the way, have you considered using Mamba for point cloud segmentation?

Hi, our serialisation strategy seems pretty suited to the Mamba model. Yet, my research interest now focuses on 3D representation learning and how to present a unified perception framework for all 3D tasks. I go back to present a new 3D backbone (PTv3) because we currently don't have a transformer good enough to go further (we still use SparseUNet in Point Prompt Training). As the issue is well solved by PTv3, I will not do some exploration on other backbone structures recently. But still, it is a good direction.

Thank you for your response, and I appreciate the excellent work you and your team have done!