NVlabs/instant-ngp
Instant neural graphics primitives: lightning fast NeRF and more
Implements four neural graphics primitives (NeRF, SDFs, neural images, and neural volumes) using multiresolution hash encoding with MLPs trained via the tiny-cuda-nn framework on NVIDIA GPUs. The architecture achieves sub-second to sub-minute training times through GPU-accelerated neural network inference and custom CUDA kernels. Includes an interactive GUI with real-time rendering, VR support, mesh conversion, camera optimization, and cross-dataset compatibility with standard NeRF formats.
17,310 stars.
Stars
17,310
Forks
2,057
Language
Cuda
License
—
Category
Last pushed
Feb 02, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/NVlabs/instant-ngp"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
nerfstudio-project/nerfstudio
A collaboration friendly studio for NeRFs
google/nerfies
This is the code for Deformable Neural Radiance Fields, a.k.a. Nerfies.
SuLvXiangXin/zipnerf-pytorch
Unofficial implementation of ZipNeRF
google/hypernerf
Code for "HyperNeRF: A Higher-Dimensional Representation for Topologically Varying Neural...
yashbhalgat/HashNeRF-pytorch
Pure PyTorch Implementation of NVIDIA paper on Instant Training of Neural Graphics primitives:...