skills.rest
yangwhale
yangwhaleCOMMUNITY
Explore all skills by this author →
GitHub

sglang-installer

Install, configure, and run SGLang inference servers on NVIDIA GPUs.

9|1|Updated Oct 20, 2025
Search Tags:#diagnostics#installation#CUDA#sglang#NVIDIA#server-start
One-click install
npx skills add https://github.com/yangwhale/gpu-tpu-pedia --skill sglang-installer
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sglang-installer
Source: https://github.com/yangwhale/gpu-tpu-pedia/tree/main/VibeCoding/claude-code/skills/sglang-installer
Command: npx skills add https://github.com/yangwhale/gpu-tpu-pedia --skill sglang-installer

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill streamlines installing SGLang from source, configuring dependencies, setting up CUDA environments, and launching a working inference server on NVIDIA GPUs (B200/H100/A100), reducing setup time and common misconfigurations.

Core Features & Use Cases

  • From-source installation and dependency management: Install SGLang and its required libraries on CUDA-enabled GPUs, handling PyTorch, sgl-kernel, FlashInfer, and NVIDIA packages.
  • Environment setup and validation: Configure CUDA_HOME, LD_LIBRARY_PATH, and related env vars; verify CUDA and PyTorch CUDA support.
  • Server startup, testing and debugging: Launch the inference server, perform health checks, and diagnose common runtime issues; supports tensor parallelism configurations and MoE readiness with DeepEP if needed.
  • Disaggregation readiness: Guidance for prefill/decode disaggregation using transfer backends like Mooncake or NIXL when deploying MoE models.

Quick Start

Clone and install SGLang from source, install NVIDIA libraries (nvidia-nccl-cu12 and nvidia-cudnn-cu12), optionally install Mooncake or NIXL for disaggregation, set up the environment, and start the server.

Frequently Asked Questions about sglang-installer

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

FAQPage Schema
How do I install and run SGLang on NVIDIA GPUs from source?▼

To install and run SGLang on NVIDIA GPUs, you need to clone the source, install dependencies like PyTorch and sgl-kernel, configure CUDA environment variables, and launch the inference server. This handles the full setup automatically.

What CUDA environment variables need to be configured for SGLang inference server setup?▼

Configuring the SGLang inference server requires setting CUDA_HOME and LD_LIBRARY_PATH environment variables. Proper configuration validates CUDA and PyTorch CUDA support to ensure the server runs correctly on your GPUs.

Can I use SGLang with tensor parallelism on H100 or A100 GPUs?▼

Yes, SGLang supports tensor parallelism configurations on NVIDIA GPUs like B200, H100, and A100. The installation process includes setting up the required NVIDIA libraries and FlashInfer to enable parallel inference.

Does SGLang support prefill and decode disaggregation for MoE models?▼

SGLang supports prefill and decode disaggregation for MoE models using transfer backends like Mooncake or NIXL. It also provides DeepEP readiness for efficient MoE deployment on your configured inference server.

Why does my SGLang server startup fail after installation?▼

SGLang server startup failures often stem from missing NVIDIA libraries or incorrect CUDA configurations. The installation process performs health checks and diagnoses common runtime issues to help identify and resolve these dependency problems.

Related Skills

Explore More Software Engineering Skills→
NVIDIA-TAONVIDIA-TAO
83

tao-run-on-docker

Run Docker containers on GPU hosts with NVIDIA Container Toolkit.

Official
Intermediate
NVIDIANVIDIA
2.8k

cuopt-server-api-python

Deploy cuOpt REST server and access routing endpoints with Python clients.

Official
Intermediate
yangwhaleyangwhale
9

vllm-installer

Install, configure, and run vLLM on NVIDIA GPUs with CUDA.

Community
Advanced
GitnappGitnapp

serving-llms-vllm

Deploy large language model inference services with vLLM and OpenAI-compatible endpoints.

Community
Advanced