Overview
- Full name
docker.io/openeuler/vllm-cpu- Registry
- docker.io
- Namespace
- openeuler
- Repository
- vllm-cpu
- Cached tags
- 16
- Last synced
- 07/05/2026, 08:18 AM
- Topics
Introduction
vLLM CPU 推理镜像（openEuler），大语言模型 CPU 推理引擎，无 GPU 环境的 LLM API 服务部署。
Details
openeuler/vllm-cpu 基于 openEuler 提供 vLLM CPU 推理版本，在无 NVIDIA GPU 环境下运行大语言模型推理。定位为 CPU LLM 推理服务，吞吐低于 GPU 版；无 GPU 服务器 LLM 部署、开发测试推理或边缘 CPU 推理场景。
快速启动
docker run -d --name vllm-cpu \
-p 8000:8000 \
-e MODEL_NAME=Qwen/Qwen2-0.5B-Instruct \
docker.io/openeuler/vllm-cpu:latest
推荐实践
docker run -d --name vllm-cpu \
--restart unless-stopped \
-p 8000:8000 \
-e MODEL_NAME=Qwen/Qwen2-0.5B-Instruct \
-v /data/models:/models \
docker.io/openeuler/vllm-cpu:0.6.0
核心参数说明
-p 8000:8000— OpenAI 兼容 APIMODEL_NAME— HuggingFace 模型 ID-v .../models:...— 模型缓存- 资源 — 大模型 CPU 推理需足够内存
Kubernetes
apiVersion: apps/v1
kind: Deployment
metadata:
name: vllm-cpu
spec:
replicas: 1
selector:
matchLabels:
app: vllm-cpu
template:
metadata:
labels:
app: vllm-cpu
spec:
containers:
- name: vllm
image: docker.io/openeuler/vllm-cpu:0.6.0
ports:
- containerPort: 8000
env:
- name: MODEL_NAME
value: Qwen/Qwen2-0.5B-Instruct
resources:
requests:
memory: 8Gi
limits:
memory: 16Gi
volumeMounts:
- name: models
mountPath: /models
volumes:
- name: models
persistentVolumeClaim:
claimName: vllm-models
---
apiVersion: v1
kind: Service
metadata:
name: vllm-cpu
spec:
selector:
app: vllm-cpu
ports:
- port: 8000
targetPort: 8000
PVC 存模型；按内存需求调 resources。
Latest tags (20)
View all 16 tags →| Tag | Arch | Pushed At | Size | Digest | Sync | Wish | Action |
|---|---|---|---|---|---|---|---|
docker.io/openeuler/vllm-cpu :latest | linux/amd64linux/arm64 | 2026/05/09 19:23 | 2.3 GB | sha256:14d8f0fee770… | |||
docker.io/openeuler/vllm-cpu :0.20.1-oe2403sp3 | linux/amd64linux/arm64 | 2026/05/09 19:23 | 2.3 GB | sha256:14d8f0fee770… | |||
docker.io/openeuler/vllm-cpu :0.18.0-oe2403sp3 | linux/amd64linux/arm64 | 2026/04/03 08:39 | 2.2 GB | sha256:b36896ded509… | |||
docker.io/openeuler/vllm-cpu :0.14.0-oe2403sp3 | linux/amd64linux/arm64 | 2026/01/27 14:57 | 2.1 GB | sha256:c24f31bd374c… | |||
docker.io/openeuler/vllm-cpu :0.10.1-oe2403lts | linux/amd64 | 2025/09/12 12:37 | 2.5 GB | sha256:ef7513c08a04… | |||
docker.io/openeuler/vllm-cpu :0.9.1-oe2403lts | linux/amd64linux/arm64 | 2025/07/04 22:47 | 3.3 GB | sha256:dd2854f06088… | |||
docker.io/openeuler/vllm-cpu :0.9.1-oe2203sp4 | linux/amd64linux/arm64 | 2025/07/04 22:23 | 3.1 GB | sha256:eebc51d65af8… | |||
docker.io/openeuler/vllm-cpu :0.9.0-oe2203sp4 | linux/amd64linux/arm64 | 2025/06/12 22:58 | 3.0 GB | sha256:4f8dc28cb4c2… | |||
docker.io/openeuler/vllm-cpu :0.8.5-oe2203sp4 | linux/amd64linux/arm64 | 2025/06/12 22:58 | 2.7 GB | sha256:3f82d309c1c7… | |||
docker.io/openeuler/vllm-cpu :0.8.4-oe2203sp4 | linux/amd64 | 2025/06/12 14:40 | 2.7 GB | sha256:243549118edb… | |||
docker.io/openeuler/vllm-cpu :0.8.3-oe2203sp4 | linux/amd64linux/arm64 | 2025/06/12 11:27 | 2.4 GB | sha256:90a2c9a6f00d… | |||
docker.io/openeuler/vllm-cpu :0.9.0-oe2403lts | linux/amd64linux/arm64 | 2025/06/10 21:27 | 3.1 GB | sha256:a6f56d768074… | |||
docker.io/openeuler/vllm-cpu :0.8.5-oe2403lts | linux/amd64linux/arm64 | 2025/06/10 21:16 | 2.8 GB | sha256:6d7c4e3d7c93… | |||
docker.io/openeuler/vllm-cpu :0.8.4-oe2403lts | linux/amd64 | 2025/06/10 21:08 | 2.8 GB | sha256:165b6a7f9aa2… | |||
docker.io/openeuler/vllm-cpu :0.8.3-oe2403lts | linux/amd64linux/arm64 | 2025/06/10 20:57 | 2.5 GB | sha256:591515b2df60… | |||
docker.io/openeuler/vllm-cpu :0.6.3-oe2403lts | linux/amd64 | 2024/10/24 16:30 | 3.0 GB | sha256:b36c99217d7c… |