This is the official implementation of our Siggrapha Asia 2024 paper "DiffCSG: Differentiable CSG via Rasterization".
Our Implementation are based on Nvdiffrast.
- Release Code and Data (Will be done before 2025, I promise.)
- Clean the code and usage (Very soon before Jan 2025.)
- Upgrade for better performance .
Install nvdiffrast under nvdiffrast_custom
Then run the benchmark under ./nvidffrast_custom:
sudo ./run_sample.sh --build-container ./samples/torch/CSGDR.py --use_random --scale_level 0 --USE_ALL_PARAM
If you are familiar with cuda, openGL and Nvdiffrast, feel free to contact me at [email protected]