sglang-gemma4-optimization

Codify Gemma 4 optimization guidance from PR diffs in SGLang.

721|65|Updated Apr 1, 2026
One-click install
npx skills add https://github.com/BBuf/AI-Infra-Auto-Driven-SKILLS --skill sglang-gemma4-optimization
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sglang-gemma4-optimization
Source: https://github.com/BBuf/AI-Infra-Auto-Driven-SKILLS/tree/main/skills/model-optimization/sglang/sglang-gemma4-optimization
Command: npx skills add https://github.com/BBuf/AI-Infra-Auto-Driven-SKILLS --skill sglang-gemma4-optimization

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The manual codifies how to audit, debug, extend, and document Gemma 4 optimization work across text, MoE, multimodal, reasoning, tool use, and quantized serving, using PR-based diffs as the production baseline.

Core Features & Use Cases

  • Non-negotiable evidence-driven workflow: enforces production-grade PR dossier standards for Gemma 4 work.
  • Cross-domain coverage: addresses Gemma 4 text, MoE, multimodal, and tool-use aspects, including CI validation and history replay.
  • Reusable decision logs: captures diffs, PR history, and validation lanes to support auditing and future updates.

Quick Start

Review the latest Gemma 4 PR diffs and apply the production-bar guidelines from skills/model-optimization/model-pr-diff-dossier/SKILL.md to update the optimization dossier.

Frequently Asked Questions about sglang-gemma4-optimization

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I optimize Gemma 4 models for text and multimodal serving in SGLang?

Gemma 4 optimization in SGLang involves auditing PR diffs to codify actionable guidance across text, MoE, multimodal, and tool-use contexts. You apply production-bar guidelines from a standard skill dossier to update optimization workflows.

What is the best way to track Gemma 4 MoE and quantized serving changes?

Tracking Gemma 4 MoE and quantized serving changes requires capturing PR diffs, history, and validation lanes into reusable decision logs. This evidence-driven workflow supports auditing and future updates for production-grade model optimization.

How do I validate Gemma 4 optimization patches using CI checks?

Validating Gemma 4 optimization patches uses CI checks replayed from PR histories to ensure code paths meet production standards. The workflow enforces non-negotiable, evidence-driven dossier standards across multiple PR baselines.

Does SGLang support tool-use and reasoning optimization for Gemma 4 models?

SGLang supports Gemma 4 tool-use and reasoning optimization by applying PR-based diffs across these contexts. The workflow addresses cross-domain coverage including tool use, reasoning, and quantized serving environments.

What prerequisites are needed to audit Gemma 4 PR diffs for SGLang?

Auditing Gemma 4 PR diffs requires access to relevant PR diffs, code paths, and validation tests. You must adhere to the production standard defined in the model-pr-diff-dossier skill manual to ensure evidence-driven optimization.

Why use PR diffs for Gemma 4 optimization instead of general documentation?

PR diffs provide a production baseline for Gemma 4 optimization, ensuring changes are evidence-driven and validated. General documentation lacks the enforced CI validation, history replay, and decision log capture inherent in the PR dossier workflow.