← All skills
vllm-ascend
75 / 100 · Review before installing
Check before installing. Some of what it can do is expected: a coding skill usually runs build and test commands.
By ascend-ai-coding/awesome-ascend-skills (GitHub) · coding · 184 stars · scanned at commit 75ec89c
vLLM Ascend plugin for LLM inference serving on Huawei Ascend NPU. Use for offline batch inference, API server deployment, quantization inference (with msmodelslim quantized models), tensor/pipeline parallelism for distributed serving, and OpenAI-compatible API endpoints. Supports Qwen, DeepSeek, GLM, LLaMA models with Ascend-optimized kernels.
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
The author does not mention this. Expected for a coding skill.
Found in 15 places.
Show 12 more
- skills/inference/vllm-ascend/SKILL.md, line 153
python -c "import vllm_ascend; print(vllm_ascend.__version__)" - skills/inference/vllm-ascend/SKILL.md, line 167
vllm serve <model_path> \ - skills/inference/vllm-ascend/SKILL.md, line 248
vllm serve /path/to/quantized-model-w8a8 \ - skills/inference/vllm-ascend/SKILL.md, line 285
vllm serve Qwen/Qwen2.5-72B-Instruct \ - skills/inference/vllm-ascend/SKILL.md, line 319
vllm serve <model> \ - skills/inference/vllm-ascend/SKILL.md, line 354
vllm serve <model> --max-model-len 4096 --max-num-seqs 256 - skills/inference/vllm-ascend/SKILL.md, line 375
npu-smi info - skills/inference/vllm-ascend/SKILL.md, line 385
vllm serve <model> --max-model-len 2048 - skills/inference/vllm-ascend/SKILL.md, line 397
ls /path/to/model - skills/inference/vllm-ascend/SKILL.md, line 409
vllm serve <model> --dtype bfloat16 - skills/inference/vllm-ascend/SKILL.md, line 421
curl http://localhost:8000/health - skills/inference/vllm-ascend/SKILL.md, line 434
export VLLM_WORKER_MULTIPROC_METHOD=spawn
Connects to the internet
The author does not mention this.
Found in 10 places.
Show 7 more
Downloads and installs other software
The author does not mention this.
Found in 15 places.
Show 12 more
- skills/inference/vllm-ascend/references/installation.md, line 144
pip install vllm-ascend==0.14.0rc1 - skills/inference/vllm-ascend/references/installation.md, line 169
pip install build wheel setuptools - skills/inference/vllm-ascend/references/installation.md, line 192
pip install -e . - skills/inference/vllm-ascend/references/installation.md, line 199
pip install -e . - skills/inference/vllm-ascend/references/installation.md, line 209
pip install -e . - skills/inference/vllm-ascend/references/installation.md, line 213
pip install -e . - skills/inference/vllm-ascend/references/installation.md, line 422
pip install vllm==0.14.1 - skills/inference/vllm-ascend/references/installation.md, line 423
pip install vllm-ascend==0.14.0rc1 - skills/inference/vllm-ascend/references/troubleshooting.md, line 32
pip install vllm==0.11.4 vllm-ascend==0.11.4 - skills/inference/vllm-ascend/references/troubleshooting.md, line 42
pip install torch-npu==2.5.1 - skills/inference/vllm-ascend/references/troubleshooting.md, line 153
pip install torch-npu==2.5.1 - skills/inference/vllm-ascend/references/troubleshooting.md, line 154
pip install vllm-ascend
Safer alternatives
Low-risk skills in the same category (coding) that look related. Check that one does what you need.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'vllm-ascend' ] || [ -L ~/.agents/skills/'vllm-ascend' ]; then echo 'A skill folder named vllm-ascend already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/ascend-ai-coding/awesome-ascend-skills' "$d" && git -C "$d" checkout --quiet '75ec89ce9d401be8ac95980e1c3f42f8d4315057' -- 'skills/inference/vllm-ascend' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/inference/vllm-ascend' ~/.agents/skills/'vllm-ascend' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'vllm-ascend' ~/.claude/skills/'vllm-ascend' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for ascend-ai-coding/awesome-ascend-skills shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/ascend-ai-coding/awesome-ascend-skills/)
Score history
| Date | Version | Score |
|---|
| 2026-11-08 | 75ec89c | 100 |