概要
- フル名
docker.io/samueltallet/alpine-llama-cpp-server- Registry
- docker.io
- 名前空間
- samueltallet
- リポジトリ
- alpine-llama-cpp-server
- キャッシュ Tag 数
- 568
- 最終同期
- 07/06/2026, 04:30 PM
- 関連トピック
概要
基于 Alpine 的 llama.cpp HTTP 推理服务镜像，加载 GGUF 模型提供 OpenAI 兼容 API，适合 CPU 边缘节点轻量 LLM 部署。
詳細
samueltallet/alpine-llama-cpp-server 启动 llama.cpp 的 server 模式，挂载 GGUF 权重后暴露 /completion 与 OpenAI 风格 /v1/chat/completions。Alpine 基础镜像体积小，默认 CPU 推理，可选 GPU 变体视 tag 而定。与 ollama 相比更底层、配置项更细；生产应固定模型 tag、只读挂载 weights，并按并发设置 --parallel 与 context 长度限制内存。
快速启动
docker run -d --name llama-server \
-p 8080:8080 \
-v $(pwd)/models:/models \
docker.io/samueltallet/alpine-llama-cpp-server:latest \
--model /models/llama-2-7b.Q4_K_M.gguf --host 0.0.0.0 --port 8080
推荐实践
docker run -d --name llama-server \
--restart unless-stopped \
-p 8080:8080 \
-v /data/models:/models:ro \
docker.io/samueltallet/alpine-llama-cpp-server:0.3.0 \
--model /models/llama-2-7b.Q4_K_M.gguf \
--host 0.0.0.0 --port 8080 --ctx-size 4096 --parallel 2
核心参数说明
-p 8080:8080— HTTP 推理 API 端口-v /data/models:/models:ro— GGUF 模型只读挂载--model— 模型文件路径--ctx-size— 上下文 token 上限--parallel— 并发推理槽位数--restart unless-stopped— 边缘推理节点持续服务
Kubernetes
apiVersion: apps/v1
kind: Deployment
metadata:
name: llama-cpp-server
spec:
replicas: 2
selector:
matchLabels:
app: llama-cpp-server
template:
metadata:
labels:
app: llama-cpp-server
spec:
containers:
- name: server
image: docker.io/samueltallet/alpine-llama-cpp-server:0.3.0
args:
- --model
- /models/llama-2-7b.Q4_K_M.gguf
- --host
- 0.0.0.0
- --port
- "8080"
ports:
- containerPort: 8080
resources:
requests:
memory: 8Gi
volumeMounts:
- name: models
mountPath: /models
readOnly: true
volumes:
- name: models
persistentVolumeClaim:
claimName: llama-models
---
apiVersion: v1
kind: Service
metadata:
name: llama-cpp-server
spec:
selector:
app: llama-cpp-server
ports:
- port: 8080
targetPort: 8080
大模型 PVC 或 initContainer 预拉取；按内存 requests 调度，GPU 变体需 nodeSelector 与 nvidia.com/gpu 资源。
最新 Tag（20 件）
すべての 568 件の Tag を見る →| Tag | アーキテクチャ | プッシュ日時 | サイズ | Digest | 同期 | 期望 | 操作 |
|---|---|---|---|---|---|---|---|
docker.io/samueltallet/alpine-llama-cpp-server :latest | linux/amd64linux/arm64 | 2026/07/01 16:35 | 12.0 MB | sha256:eb37eda5d06d… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9853 | linux/amd64linux/arm64 | 2026/07/01 16:35 | 12.0 MB | sha256:eb37eda5d06d… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9843 | linux/amd64linux/arm64 | 2026/06/30 16:33 | 12.0 MB | sha256:98d36d97f625… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9837 | linux/amd64linux/arm64 | 2026/06/29 16:37 | 11.9 MB | sha256:d20bffadc461… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9828 | linux/amd64linux/arm64 | 2026/06/28 16:34 | 11.9 MB | sha256:c9d1be664d62… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9821 | linux/amd64linux/arm64 | 2026/06/27 15:48 | 12.0 MB | sha256:afb87d7b0232… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9803 | linux/amd64linux/arm64 | 2026/06/26 16:27 | 11.9 MB | sha256:66a0239dc107… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9787 | linux/amd64linux/arm64 | 2026/06/25 16:15 | 11.9 MB | sha256:9caae0175745… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9776 | linux/amd64linux/arm64 | 2026/06/24 16:26 | 11.9 MB | sha256:adfb397b4208… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9765 | linux/amd64linux/arm64 | 2026/06/23 16:33 | 11.9 MB | sha256:29ee07302ca9… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9755 | linux/amd64linux/arm64 | 2026/06/22 16:49 | 11.9 MB | sha256:a2039aeeb111… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9744 | linux/amd64linux/arm64 | 2026/06/21 16:35 | 11.8 MB | sha256:4e31f4b476b6… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9733 | linux/amd64linux/arm64 | 2026/06/20 16:34 | 11.8 MB | sha256:fd354666415f… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9717 | linux/amd64linux/arm64 | 2026/06/19 16:49 | 11.8 MB | sha256:2096adf1d628… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9692 | linux/amd64linux/arm64 | 2026/06/18 16:39 | 11.8 MB | sha256:d64af22f339d… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9672 | linux/amd64linux/arm64 | 2026/06/17 16:17 | 11.9 MB | sha256:6ebc9125f2da… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9664 | linux/amd64linux/arm64 | 2026/06/16 16:45 | 11.8 MB | sha256:fb0db850aa2e… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9641 | linux/amd64linux/arm64 | 2026/06/15 16:43 | 11.9 MB | sha256:5c59d43e17b8… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9628 | linux/amd64linux/arm64 | 2026/06/14 16:30 | 11.8 MB | sha256:2e09a551617d… | |||
docker.io/samueltallet/alpine-llama-cpp-server :b9603 | linux/amd64linux/arm64 | 2026/06/12 16:37 | 11.7 MB | sha256:cd693fd8e208… |