Sglang Sota H200 Prompts banner
BBuf BBuf

Sglang Sota H200 Prompts

Development community

Description

# H200 prompts ## 1 GPU ### nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 ```text 使用 sglang-sota-humanize-loop skill。 model_id: nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 root_dir: /Users/bbuf/工作目录/Common target_hardware: single-node 1x NVIDIA H200 minimum_gpu_count: 1 precision_quantization: FP8 initial_deployment: SGLang TP=1 要求: 远端使用 ion8-h200 或 ion9-h200;SGLang 使用已有 sglang_bbuf 容器,容器内 repo 为 /home/sglang-omni/bbuf/repos/sglang。 要求: vLLM 和 TensorRT-LLM 直接使用最新镜像 vllm/vllm-openai:latest 与 nvcr.io/

Installation

Installs to ~/.claude/skills/bbuf-ai-infra-auto-driven-skills-sglang-sota-h200-prompts/SKILL.md

Terminal
mkdir -p ~/.claude/skills/bbuf-ai-infra-auto-driven-skills-sglang-sota-h200-prompts && curl -fsSL https://raw.githubusercontent.com/BBuf/AI-Infra-Auto-Driven-SKILLS/HEAD/prompts/sglang-sota-h200-prompts.md -o ~/.claude/skills/bbuf-ai-infra-auto-driven-skills-sglang-sota-h200-prompts/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.

Full documentation available on GitHub

View Source Repository