Overview
- Full name
docker.io/approachingai/ktransformers- Registry
- docker.io
- Namespace
- approachingai
- Repository
- ktransformers
- Cached tags
- 50
- Last synced
- 07/13/2026, 09:39 AM
Introduction
KTransformers 大模型推理优化镜像,CPU/GPU 混合推理加速,降低 LLM serving 硬件门槛,需 GPU 与模型缓存。
Details
approachingai/ktransformers 运行 KTransformers 大模型推理优化框架,通过 CPU/GPU 混合计算降低 LLM serving 显存需求。定位为 GPU/CPU 混合 LLM 推理服务,需 GPU 与模型缓存 volume;docker GPU 节点部署,非 CPU-only K8s 常规场景。
快速启动
docker run -d --name ktransformers \
--gpus all \
-p 8080:8080 \
docker.io/approachingai/ktransformers:latest
推荐实践
docker run -d --name ktransformers \
--gpus all \
--restart unless-stopped \
-p 8080:8080 \
-v /data/models:/models \
docker.io/approachingai/ktransformers:0.2.0
核心参数说明
--gpus all— GPU 推理加速-p 8080:8080— 推理 REST API-v .../models:...— 模型权重缓存- 定位 — 混合 CPU/GPU LLM serving
Latest tags (20)
View all 50 tags →| Tag | Arch | Pushed At | Size | Digest | Sync | Wish | Action |
|---|---|---|---|---|---|---|---|
docker.io/approachingai/ktransformers :v0.5.3 | linux/amd64 | 2026/04/02 20:36 | 14.5 GB | sha256:06c6aee9e50b… | |||
docker.io/approachingai/ktransformers :latest | linux/amd64 | 2026/03/05 19:20 | 14.5 GB | sha256:8476770c0693… | |||
docker.io/approachingai/ktransformers :v0.5.2.post2 | linux/amd64 | 2026/03/05 19:14 | 14.5 GB | sha256:8476770c0693… | |||
docker.io/approachingai/ktransformers :v0.4.3-cu128 | linux/amd64 | 2025/12/11 14:01 | 21.4 GB | sha256:88de78c03f0e… | |||
docker.io/approachingai/ktransformers :v0.3.2-FANCY | linux/amd64 | 2025/07/01 14:07 | 8.1 GB | sha256:2841a385ac1e… | |||
docker.io/approachingai/ktransformers :latest-FANCY | linux/amd64 | 2025/07/01 14:07 | 8.1 GB | sha256:2841a385ac1e… | |||
docker.io/approachingai/ktransformers :v0.3.2-AVX512 | linux/amd64 | 2025/07/01 14:05 | 8.1 GB | sha256:47fe7aa75a6a… | |||
docker.io/approachingai/ktransformers :latest-AVX512 | linux/amd64 | 2025/07/01 14:05 | 8.1 GB | sha256:47fe7aa75a6a… | |||
docker.io/approachingai/ktransformers :v0.3.2-AVX2 | linux/amd64 | 2025/07/01 14:03 | 8.1 GB | sha256:cb9e708c81eb… | |||
docker.io/approachingai/ktransformers :latest-AVX2 | linux/amd64 | 2025/07/01 14:03 | 8.1 GB | sha256:cb9e708c81eb… | |||
docker.io/approachingai/ktransformers :v0.3.2-NATIVE | linux/amd64 | 2025/07/01 14:02 | 8.1 GB | sha256:28b98699dd8d… | |||
docker.io/approachingai/ktransformers :latest-NATIVE | linux/amd64 | 2025/07/01 14:02 | 8.1 GB | sha256:28b98699dd8d… | |||
docker.io/approachingai/ktransformers :v0.3.1-AVX512 | linux/amd64 | 2025/05/17 18:16 | 7.8 GB | sha256:3e4aa0ac62fa… | |||
docker.io/approachingai/ktransformers :v0.3.1-NATIVE | linux/arm64 | 2025/05/17 18:12 | 7.8 GB | sha256:dbcb5d2dcca7… | |||
docker.io/approachingai/ktransformers :v0.3.1-FANCY | linux/amd64 | 2025/05/17 18:07 | 7.8 GB | sha256:b1dc1d9a1f62… | |||
docker.io/approachingai/ktransformers :v0.3.1-AVX2 | linux/amd64 | 2025/05/17 18:06 | 7.8 GB | sha256:c779be2b70d4… | |||
docker.io/approachingai/ktransformers :v0.3-AVX2 | linux/amd64 | 2025/04/29 10:04 | 11.9 GB | sha256:b630eedee857… | |||
docker.io/approachingai/ktransformers :v0.3-NATIVE | linux/amd64 | 2025/04/29 09:55 | 11.9 GB | sha256:44fa51d805e6… | |||
docker.io/approachingai/ktransformers :v0.3-FANCY | linux/amd64 | 2025/04/29 09:47 | 11.9 GB | sha256:14d3c8e8c1e4… | |||
docker.io/approachingai/ktransformers :v0.3-AVX512 | linux/amd64 | 2025/04/29 09:47 | 11.9 GB | sha256:d0d2d6eb7077… |