about.md
examples/demo-corpus/data/research/repos/nvlabs-instant-ngp/about.md
About NVlabs/instant-ngp
NVIDIA's reference implementation of Instant Neural Graphics Primitives with a Multiresolution Hash Encoding (Müller et al., 2022). See the paper page for context.
Introduces the multiresolution hash grid encoding and tiny-CUDA-NN fused MLP runtime that together brought NeRF training time from hours to seconds, and the Instant-NGP viewer that made interactive radiance-field editing practical. The encoding is reused by MERF, Co-SLAM, NICE-SLAM, and many SLAM/4D follow-ups in the corpus. The upstream license is NVIDIA non-commercial research only, so this corpus mirrors only the README and explicitly omits the code.