Skills
.Work. You
Rest

vllm-deepseek-v32-optimization

Community

PR-driven optimization for vLLM DeepSeek V3.2.

Software Engineering#optimization#runtime#deepseek#vllm#indexer#pr-diff-audit#tool-parser
AuthorBBuf
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This skill provides a production-grade guide for auditing, debugging, extending, and documenting the DeepSeek V3.2 integration within vLLM. It focuses on the optimization work driven by PR diffs across the runtime, indexer, tool parser, MTP fallback, and long-context decode kernels.

Core Features & Use Cases

  • PR-diff audit framework: captures, categorizes, and references DeepSeek V3.2 changes across the vLLM stack.
  • Traceable history integration: links to PR notes and history references to support reproducibility and onboarding.
  • Implementation guidance: synthesizes diffs into actionable tasks for runtime, indexing, tool parsing, and decode kernels.
  • Use Case: A dev lead uses this to align the vLLM DeepSeek V3.2 integration with the current mainline and ROCm/AMD PR considerations.

Quick Start

Review the DeepSeek V3.2 PR history and apply the recommended optimization steps to integrate and validate the changes in your vLLM deployment.

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-deepseek-v32-optimization
Download link: https://github.com/BBuf/AI-Infra-Auto-Driven-SKILLS/archive/main.zip#vllm-deepseek-v32-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.