README.md

---
license: mit
tags:
- point-cloud
- 3d-semantic-segmentation
- mixture-of-experts
---

# Point-MoE: Large-Scale Multi-Dataset Training with Mixture-of-Experts for 3D Semantic Segmentation

[Paper (arXiv 2505.23926)](https://arxiv.org/abs/2505.23926) · [Project Page](https://point-moe.cs.virginia.edu/) · [Code](https://github.com/UVA-Computer-Vision-Lab/point_moe) · [Google Drive mirror](https://drive.google.com/drive/folders/1myrNSuVBmSfI2xXv69RMXrWjE-W9DEyQ?usp=drive_link)

Pretrained Point-MoE-L checkpoints for the paper *Point-MoE: Large-Scale Multi-Dataset Training with Mixture-of-Experts for 3D Semantic Segmentation* (ICLR 2026).

## Checkpoints

| Checkpoint | Config | Training data |
|---|---|---|
| `pointmoe_l_indoor_last.pth` | `pointmoe_l_indoor_last.py` | ScanNet, Structured3D, S3DIS (Area 1/2/3/4/6) |
| `pointmoe_l_indoor_outdoor_last.pth` | `pointmoe_l_indoor_outdoor_last.py` | ScanNet, Structured3D, S3DIS (Area 1/2/3/4/6), SemanticKITTI, nuScenes |

All models are trained on the official training splits only.

**Model:** Point Transformer V3 backbone with Mixture-of-Experts layers (8 experts, top-2 routing). ~100M total parameters, ~59M activated per token.

## Usage

1. Set up the environment and prepare the datasets under `data/` following the [code repository](https://github.com/UVA-Computer-Vision-Lab/point_moe).
2. Download a checkpoint and its config:

```bash
huggingface-cli download uva-cv-lab/Point_MoE pointmoe_l_indoor_last.pth pointmoe_l_indoor_last.py --local-dir checkpoints
```

3. Evaluate from the root of the code repository:

```bash
python tools/train.py \
  --config-file checkpoints/pointmoe_l_indoor_last.py \
  --num-gpus 4 \
  --options weight=checkpoints/pointmoe_l_indoor_last.pth save_path=exp/pointmoe_l_indoor
```

## Citation

```bibtex
@inproceedings{chenpoint,
  title={Point-MoE: Large-Scale Multi-Dataset Training with Mixture-of-Experts for 3D Semantic Segmentation},
  author={Chen, Xuweiyi and Zhou, Wentao and RoyChowdhury, Aruni and Cheng, Zezhou},
  booktitle={The Fourteenth International Conference on Learning Representations},
  year={2026}
}
```

## Acknowledgements

Built on [Pointcept](https://github.com/Pointcept/Pointcept).