sxyu/pixel-nerf

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1465 stars · 205 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Official PyTorch implementation of pixelNeRF, a CVPR 2021 method that predicts neural radiance fields conditioned on one or few input images for feed-forward novel view synthesis. It includes training and evaluation code, pretrained checkpoints, and dataset adapters for ShapeNet, DTU, and SRN benchmarks.

Use cases

• synthesize novel views of an object from a single image
• reconstruct 3D scene representation from sparse views without test-time optimization
• train a NeRF model across multiple ShapeNet categories
• evaluate single-image view synthesis on DTU and SRN benchmarks
• render custom datasets for NeRF-style training

When to avoid

• you need per-scene high-fidelity NeRF reconstructions with many views
• you want a production-ready or actively maintained pipeline
• you need the multi-object experiment, which is not yet implemented

Facets

library · maturity maintenance

machine-learning deep-learning image-processing graphics simulation computer-vision machine-learning graphics python neural-radiance-fields nerf novel-view-synthesis 3d-reconstruction cvpr-2021 pytorch research-code research linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="sxyu/pixel-nerf")