Installation • Usage • Citation • Concepts
Pointelligence is a high-performance library for 3D point cloud deep learning research. It provides efficient GPU-accelerated primitives and ready-to-use neural network architectures for spatial intelligence tasks.
| Feature | Description |
|---|---|
| 🎯 PointCNN++ | Official implementation of PointCNN++ (CVPR 2026) — a significant evolution of PointCNN (NeurIPS 2018) |
| ⚡ High Performance | Optimized CUDA kernels for native point convolution with minimal memory overhead |
| 📦 Ragged Tensors | Efficient batching without padding — process only valid data |
| 🔧 Modular Design | Build custom architectures from composable primitives |
| 🐳 Docker Ready | One-command setup with pre-built CUDA extensions |
PointCNN++ delivers state-of-the-art performance with significantly lower memory usage and faster training times compared to existing methods.
Our native point-based approach fundamentally avoids the overhead of voxel-based auxiliary data structures:
|
|
|
Our custom Triton kernels (MVMR for forward, VVOR for backward) provide exceptional speed in both inference and training:
|
|
|
Clone the repository with third-party submodules (FCGF and Pointcept) recursively:
git clone --recursive https://github.com/ant-research/pointelligence.git
cd pointelligenceFor reproducibility, checkout the following commits in the submodules:
# FCGF (examples/FCGF)
cd examples/FCGF && git checkout pointcnnpp-version && cd ../..
# Pointcept (examples/Pointcept)
cd examples/Pointcept && git checkout pointcnnpp-version && cd ../..If you have already cloned without --recursive, run git submodule update --init --recursive to fetch the submodules.
Some operators are implemented with C++/CUDA as PyTorch extensions, which could be built and installed with the following commands:
conda create -n pointelligence python=3.10 -y
conda activate pointelligence
pip install -r requirements.txt
cd extensions
pip install --no-build-isolation -e .Use Docker for a containerized environment with all dependencies pre-installed:
# Build the Docker image
docker build -t pointelligence .
# Test the containerized environment
docker run --gpus all -it -v $(pwd):/workspace pointelligence
# Verify installation
python -m pytest tests/unittest/ -vThe Docker image includes:
- CUDA 12.4 + cuDNN + PyTorch 2.6.0+ with GPU support
- Pre-built CUDA extensions (
sparse_engines_cuda) - All system dependencies and Python packages
- Sample data preloaded
- Ready-to-use development environment
See examples/FCGF for a full training pipeline using Fully Convolutional Geometric Features with PointCNN++ as the backbone.
See examples/Pointcept for semantic segmentation using the Pointcept framework with PointCNN++ integration.
Pointelligence is the repo for the official implementation of:
- PointCNN++: Performant Convolution on Native Points
Lihan Li, Haofeng Zhong, Rui Bu, Mingchao Sun, Wenzheng Chen, Baoquan Chen, Yangyan Li@misc{li2025pointcnnperformantconvolutionnative, title={PointCNN++: Performant Convolution on Native Points}, author={Lihan Li and Haofeng Zhong and Rui Bu and Mingchao Sun and Wenzheng Chen and Baoquan Chen and Yangyan Li}, year={2025}, eprint={2511.23227}, archivePrefix={arXiv}, primaryClass={cs.CV}, url={https://arxiv.org/abs/2511.23227}, }
To ensure they are tracked effectively, please submit feature requests and issue reports here rather than via email.
For building custom architectures, see docs/ADVANCED.md covering:
- Ragged tensors — efficient batching without padding
- Neighborhoods — fixed-radius search producing (i, j) pairs
- Convolution triplets — extending (i, j) to (i, j, k) to route data through kernel weights
- MVMR — the sparse convolution operator:
output[i] += weight[k] @ input[j]