SGLang Summit talk details battle-tested NVFP4 4-bit RL across thousands of GPUs
ericzelikman · x · 2026-11-14
At SGLang Summit, Ziang Li of humans& presented production low-precision RL: 4-bit (NVFP4) training and W4A16 decode, with benchmarks to follow in a July write-up.
- Techniques battle-tested on large-scale, long-horizon RL over thousands of GPUs and trillions of tokens
- The team says it will contribute the work back to the open-source community this year, crediting SGLang, RadixArk and NVIDIA
- Their blog humans&.ai includes related pieces like "The 4-bitter Lesson" on balancing stability and performance in NVFP4 RL
More from Infra
- Running DeepSeek 4.1 Flash on a 32GB Mac via SSD streaming with open-source Local Code — alexcovo_eth · 2026-11-14
- Will local models keep up with frontier? Users debate a hardware-bound gap — AshAtlasDev · 2026-11-14
- Anthropic IPO roadshow slide leak sparks Blackwell training speculation — and a block — teortaxesTex · 2026-11-14
- New Yorker interview: AI is a brute-force problem that will 'cook the planet' — kevinnbass · 2026-11-14
- Lenore unveils hybrid Token factory in Zhangjiakou: 15,000P cluster, 2 quadrillion tokens a year — 量子位 · 2026-11-14
- LexiSuite: a Rust-built inference engine beats llama.cpp 1.0.0 at 262k context — W61k3r · 2026-11-14