< cd ../gallery

[DevTools]

GGUF-to-NVFP4-SM120 by lna-lab

go to website ↗github

Lna-Lab production pipeline: GGUF -> modelopt-format NVFP4 + working MTP head for vLLM on RTX PRO 6000 Blackwell (SM120). Stages 2 (NVFP4) and 3 (MTP graft) are Lna-Lab originals; stage 1 (GGUF->bf16) reuses li-yifei/gguf-to-nvfp4. Maintained by lna-lab on GitHub, where it has earned 10 stars from the community.

maintainer

★ 10 stars

# install

pip install gguf-to-nvfp4-sm120

# tags

Python Blackwell Gguf Modelopt Mtp

languages

last commit5mo ago

license—

more vLLM repos

Reproducible recipe: serve abliterated Gemma-4-12B (gemma4_unified) at 50-118 tok/s on no-NVLink Blackwell (SM120) via vLLM nightly + ModelOpt FP8/NVFP4 + MTP spec-decode. Maintained by lna-lab on GitHub, where it has earned 15 stars from the community.

@lna-lab4mo ago

DGX Spark research and tests - containers, benchmarks, and investigation notes for running models on GB10 (SM 12.1). Maintained by Sggin1 on GitHub, where it has earned 18 stars from the community.

GLM-5.2 744B with a true 4-bit NVFP4 KV cache on 4x DGX Spark: 42 tok/s peak, 317K-token KV pool (+58.6% vs fp8), needle-verified at 250K depth, serving 316K context. Maintained by tonyd2wild on GitHub, where it has earned 12 stars from the community.

@tonyd2wild3mo ago