lna-lab/gemma4-12b-vllm-sm120

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.

14
Stars
0
Forks
17.6
Radar Score
Stable
Momentum

Scoring Breakdown

Impact2.5
Velocity24.0
Health32.7

Growth Metrics

7-Day Stars
+0
30-Day Stars
+0
Acceleration
x
Language
Python

Why it ranks

Modest but positive momentum.

LLM Frameworks ยท inference engineTop 99% globally
View on GitHub