# ashawkey/torch-ngp

A pytorch CUDA extension implementation of instant-ngp (sdf and nerf), with a GUI.

Repository: https://github.com/ashawkey/torch-ngp
Canonical: https://ross.abutalabs.com/products/torch-ngp
Language: Python
License: MIT
License Family: permissive
Topics: pytorch, instant-ngp, nerf, real-time, gui, sdf, tensorf, d-nerf
Last push: 2023-11-10T03:06:57+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1685, "days_push": 1027, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2208, forks 286 (observed 2026-08-28T04:06:26.141411+00:00)

## What it is
A PyTorch implementation of instant-ngp's NeRF and SDF components with CUDA extensions and an interactive GUI for training and visualization. It also includes implementations of TensoRF, CCNeRF, and D-NeRF adapted to the instant-ngp framework.

## Use cases
- train a NeRF model from images in PyTorch
- reconstruct 3D scenes with instant-ngp hash encoding
- visualize NeRF training interactively with a GUI
- implement signed distance fields with neural networks
- model dynamic scenes with D-NeRF
- experiment with TensoRF tensorial radiance fields
- generate novel views of real scenes from photos

## When to choose
- you want a readable PyTorch codebase for NeRF research instead of the original CUDA instant-ngp
- you need an interactive GUI for training and rendering radiance fields
- you want to experiment with NeRF variants like D-NeRF, TensoRF, or composable NeRF in one framework

## When to avoid
- you need maximum training/rendering performance of the original NVIDIA instant-ngp implementation
- you want a production-ready, actively maintained NeRF pipeline - the author recommends nerf_template instead
- you need a pure PyTorch solution without CUDA extension compilation

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, graphics, gui, gpu-computing
- domain: computer-vision, graphics, machine-learning, deep-learning
- platform: python, windows
- tags: nerf, instant-ngp, sdf, neural-radiance-fields, pytorch, cuda-extension, tensorf, d-nerf, 3d-reconstruction, real-time-rendering, linux, gpu

## Member repositories
- ashawkey/torch-ngp (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:26.141411+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:46:20.236365+00:00, confidence not recorded.
  - readme: https://github.com/ashawkey/torch-ngp (fetched 2026-08-28T04:06:26.141411+00:00, sha 5087565fdd2d)
- Data as of 2026-08-30T08:39:29.467469+00:00.