LitePT: Lighter Yet Stronger Point Transformer

Yuanwen Yue, Damien Robert, Jianyuan Wang, Sunghwan Hong, Jan Dirk Wegner, Christian Rupprecht, Konrad Schindler · CVPR 2026 · 2026

arXiv Code Project page

Status

To read

Topics

backbones

Feeds

First paper · height sampling

What they did

Finds that convolution suits early, high-resolution layers and attention suits deep, low-resolution layers, and builds LitePT, a lighter point transformer, on that principle.

My note

No note yet.

Code

Repository

prs-eth/LitePT

Framework

PyTorch

Licence

MIT

My status

Not tried

BibTeX

@inproceedings{yue2026-litept,
  title = {{LitePT: Lighter Yet Stronger Point Transformer}},
  author = {Yuanwen Yue and Damien Robert and Jianyuan Wang and Sunghwan Hong and Jan Dirk Wegner and Christian Rupprecht and Konrad Schindler},
  booktitle = {CVPR 2026},
  year = {2026},
  eprint = {2512.13689},
  archivePrefix = {arXiv}
}