skills.rest
nv-action
nv-actionOFFICIAL
Explore all skills by this author →
GitHub

vllm-ascend-model-adapter

Adapt and validate models for vLLM Ascend NPU with a signed commit.

4|6|Updated Feb 28, 2025
Search Tags:#deployment#inference#vllm#ascend#npu#model-adapter
One-click install
npx skills add https://github.com/nv-action/vllm-benchmarks --skill vllm-ascend-model-adapter-nv-action
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vllm-ascend-model-adapter
Source: https://github.com/nv-action/vllm-benchmarks/tree/main/.agents/skills/vllm-ascend-model-adapter
Command: npx skills add https://github.com/nv-action/vllm-benchmarks --skill vllm-ascend-model-adapter-nv-action

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Adapt and debug existing or new models for vLLM on Ascend NPU. Implement in /vllm-workspace/vllm and /vllm-workspace/vllm-ascend, validate via direct vllm serve from /workspace, and deliver one signed commit in the current repo.

Core Features & Use Cases

  • Adapt Hugging Face or local models to run on vLLM Ascend with minimal changes.
  • Deterministic validation via startup/inference smoke tests, ACLGraph/MTP/EP as applicable.
  • Single-commit handoff delivering the final adapter to the target repository.

Quick Start

Run the vLLM Ascend model adapter workflow from /workspace, validate via direct vLLM serve, and deliver a single signed commit to the current repository.

Frequently Asked Questions about vllm-ascend-model-adapter

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

FAQPage Schema
How do I adapt a Hugging Face model to run on vLLM Ascend NPU?▼

To adapt a Hugging Face model to vLLM Ascend, implement the adapter under /vllm-workspace/vllm and /vllm-workspace/vllm-ascend, then validate it via direct vLLM serve from /workspace.

What is the validation process for a vLLM Ascend model adapter?▼

Validation for a vLLM Ascend model adapter involves deterministic startup and inference smoke tests, along with checking ACLGraph, MTP, or EP features where applicable.

Can I use vLLM for inference on Ascend NPU with local models?▼

Yes, you can adapt local models to run on vLLM Ascend with minimal changes, validating the deployment through direct vLLM serve from the /workspace directory.

How do I deliver a vLLM Ascend model adapter to a target repository?▼

Deliver a vLLM Ascend model adapter by creating a single signed commit in the current repository, ensuring a clean one-commit handoff of the final validated code.

Does the vLLM Ascend adapter workflow support Mixture of Experts features?▼

Yes, the vLLM Ascend adapter workflow supports MoE features where applicable, validating them alongside standard startup and inference checks during the adaptation process.

Related Skills

Explore More Software Engineering Skills→
Alex-stack-hubAlex-stack-hub

vllm-ascend-model-adapter

Adapt Hugging Face or local models for vLLM on Ascend NPU.

Community
Advanced
pmcdowallpmcdowall

serving-llms-vllm

Configure vLLM deployments with OpenAI-compatible APIs and performance tuning.

Community
Advanced
vllm-projectvllm-project
2.6k

vllm-ascend-model-adapter

Adapts ML models for vLLM-Ascend NPUs with deterministic validation and signing.

Official
Advanced
adaptyvbioadaptyvbio
151

protein-design-workflow

Plan protein design workflows from target preparation to validation.

Official
Advanced