vllm-nemotron-super-optimization

Community

PR-driven Nemotron optimization guide in vLLM.

AuthorBBuf
Version1.0.0
Installs0

System Documentation

What problem does it solve?

Identify and apply production-grade optimizations to Nemotron family models within vLLM, emphasizing diff-based evidence and reproducible improvements.

Core Features & Use Cases

  • PR-diff driven optimization workflow for NemotronH, Nemotron-3-Super, Nemotron Nano VL, and Nemotron VL architectures.
  • Guidance covers model wrappers, MoE, MTP, NVFP4, and VL adjacencies, with references to runtime surfaces and PR histories.
  • Validation paths include startup checks, quantization, and end-to-end tests; uses the Non-Negotiable Evidence Rule as the standard.

Quick Start

Review an open Nemotron optimization PR and apply the diff-driven changes to NemotronH or Nemotron Nano VL in vLLM, then run the corresponding validation lanes.

Dependency Matrix

Required Modules

None required

Components

references

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: vllm-nemotron-super-optimization
Download link: https://github.com/BBuf/AI-Infra-Auto-Driven-SKILLS/archive/main.zip#vllm-nemotron-super-optimization

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.