Fused Zip-NeRF: CUDA rewrite trains 6-7x faster, renders 10x faster on a 5090
jon_barron · x · 2026-10-31
fhahlbohm built a fully fused CUDA implementation of Zip-NeRF (with Claude Opus's help): 36-minute training and 300ms per megapixel rendering on an RTX 5090, matching official quality while training 6-7x and rendering 10x faster. Jon Barron, Zip-NeRF's original author, shared it saying "NeRF is back" and betting that optimized NeRF-like raytracing will eventually compete with rasterized graphics primitives. MIT-licensed on GitHub.
More from Research
- Researcher: capability transfer hits a ceiling when training against less human-like models — saurabh_shah2 · 2026-10-31
- Compaction study: agent constraints vanish 6x more often than task goals in long sessions — maksym_andr · 2026-10-31
- CMU's Gym-Anything uses coding agents to auto-build 200 robot RL environments; best model scores 0% — AkariAsai · 2026-10-31
- Cherimoya ChIP-seq models ship as-is: bad underlying data means bad predictions — jmschreiber91 · 2026-10-31
- When chromatin models disagree: null accessibility shift but a disrupted CEBPG binding site — jmschreiber91 · 2026-10-31
- Whole-Body Dexterity turns the humanoid body into one large dexterous manipulator — chris_j_paxton · 2026-10-31