vllm-project/vllm-ascend
Hardware plugin for running vLLM on Huawei Ascend NPU
- Chart
- Not on a chart right now
- Stars
- 3.1k
What it is
A community-maintained hardware plugin that enables vLLM to run on Huawei Ascend NPUs. It follows vLLM's hardware-pluggable interface to decouple Ascend NPU integration.
Who it is for
AI engineers and infrastructure teams deploying or serving open-source LLMs with vLLM on Huawei Ascend hardware.
Why it is trending
It recently released version v0.30.0 in late October 2026 and has accumulated over 3,100 stars. It serves as the vLLM community's recommended approach for supporting the Ascend backend.
Main features
- Supports Transformer, MoE, Embedding, and multi-modal LLM models
- Decouples Ascend NPU integration via vLLM's hardware-pluggable interface
- Compatible with Atlas 800I A2/A3, Atlas 300I Duo, and 950PR/950DT hardware
- Maintains versioned release branches aligned with vLLM and Ascend CI
Summarized on November 18, 2026 from the README and repository data. It can be wrong; check the source before relying on it. Report a wrong summary