nerfstudio-project/nerfacc
A PyTorch acceleration toolbox for Neural Radiance Fields, focusing on efficient volumetric sampling during training and inference.
Velocity · 7d
+1.1
★ / day
Trend
→steady
star history
NerfAcc provides a plug-and-play acceleration library for NeRF-based radiance field rendering. It optimizes the volumetric rendering pipeline by improving sampling efficiency, achieving significant speedups in training various NeRF models with minimal code modifications. The toolbox supports JIT compilation of CUDA kernels and offers a pure Python interface.