vistart/cuda Docker Image Overview

vistart/cuda

vistart

CUDA CUDNN NCCL TensorRT

17 次收藏下载次数: 0状态：社区镜像维护者：vistart仓库类型：镜像最近更新：4 年前

轩辕镜像，加速的不只是镜像。点击查看

镜像简介版本下载

轩辕镜像，加速的不只是镜像。点击查看

WARNING!!!

All images inherited from ubuntu of Official Version are confusing, PLEASE DO NOT UPDATE! We are now re-build all images from source.

![Build Status]([***]

Introduction

All images under this repository are built in strict accordance with the Dockerfiles published by NVIDIA respectively. The only difference between these images and the official version is that the operating system source is from Tsinghua University. These images are updated weekly.

In addition to storing all images in Docker Hub, they are also stored in the Hong Kong node of Aliyun Container Service for alternative. If you want to use the latest image as soon as possible, but Docker Hub has not been able to synchronize for a long time, you can try to use the Hong Kong node of Aliyun Container Service.

The tag name format is:

registry.cn-hongkong.aliyuncs.com/vistart_public/cuda:<tag name>

If you want to use the latest image as soon as possible, but don't want to change it's tag name. You can pull the image stored in the Docker Hub after pulling it stored in the Hong Kong node. Because the contents of the two are exactly the same, they have been tagged after pulling the manifest of the image stored in Docker Hub without pulling any image layer.

After that, if you want to update the image, repeat the above operations.

Latest CUDA 11.0 (Experimental)(CUDA 11.0 for Ubuntu 20.04 is Ready! CUDNN 8.0.2/NCCL 2.7.8/TensorRT 7.1.3 are coming!)

Ubuntu 20.04

11.0-base-ubuntu20.04 (cuda 11.0.3)
11.0-runtime-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8)
11.0-devel-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8)
11.0-cudnn8-runtime-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-devel-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-tensorrt7-devel-ubuntu20.04, 11.0-ubuntu20.04, latest (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 18.04

11.0-base-ubuntu18.04 (cuda 11.0.3)
11.0-runtime-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8)
11.0-devel-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8)
11.0-cudnn8-runtime-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-devel-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-tensorrt7-devel-ubuntu18.04, 10.2-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 16.04

11.0-base-ubuntu16.04 (cuda 11.0.3)
11.0-runtime-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8)
11.0-devel-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8)
11.0-cudnn8-runtime-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-devel-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-tensorrt7-devel-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

CUDA 10.2

Ubuntu 20.04 (Deprecated)

10.2-base-ubuntu20.04 (cuda 10.2.89)
10.2-runtime-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-runtime-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn7-tensorrt6-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.8)
10.2-cudnn7-tensorrt7-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 7.0.0.11)
10.2-cudnn8-runtime-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39)
10.2-cudnn8-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39)
10.2-cudnn8-tensorrt7-devel-ubuntu20.04, 10.2-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 18.04

10.2-base-ubuntu18.04 (cuda 10.2.89)
10.2-runtime-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-runtime-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn7-tensorrt6-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.8)
10.2-cudnn7-tensorrt7-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 7.0.0.11)
10.2-cudnn8-runtime-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-tensorrt7-devel-ubuntu18.04, 10.2-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 16.04

10.2-base-ubuntu16.04 (cuda 10.2.89)
10.2-runtime-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-runtime-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn7-tensorrt6-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.8)
10.2-cudnn7-tensorrt7-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 7.0.0.11)
10.2-cudnn8-runtime-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-tensorrt7-devel-ubuntu16.04, 10.2-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

CUDA 10.1

Ubuntu 20.04 (Deprecated)

10.1-base-ubuntu20.04 (cuda 10.1.243)
10.1-runtime-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-devel-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn7-tensorrt6-devel-ubuntu20.04, 10.1-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Ubuntu 18.04

10.1-base-ubuntu18.04 (cuda 10.1.243)
10.1-runtime-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-runtime-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-devel-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-devel-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-cudnn7-tensorrt6-devel-ubuntu18.04, 10.1-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Ubuntu 16.04

10.1-base-ubuntu16.04 (cuda 10.1.243)
10.1-runtime-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-runtime-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-devel-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-devel-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-cudnn7-tensorrt6-devel-ubuntu16.04, 10.1-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Ubuntu 14.04

10.1-base-ubuntu14.04 (cuda 10.1.243)
10.1-runtime-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-devel-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn7-tensorrt6-devel-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32 tensorrt 6.0.1.5)

CUDA 10.0

Ubuntu 20.04 (Deprecated)

10.0-base-ubuntu20.04 (cuda 10.0.130)
10.0-runtime-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32 tensorrt 6.0.1.8)
10.0-cudnn7-tensorrt7-devel-ubuntu20.04, 10.0-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 7.0.0.11)

Ubuntu 18.04 (No longer updated)

10.0-base-ubuntu18.04 (cuda 10.0.130)
10.0-runtime-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 6.0.1.5)
10.0-cudnn7-tensorrt7-devel-ubuntu18.04, 10.0-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 7.0.0.11)

Ubuntu 16.04 (No longer updated)

10.0-base-ubuntu16.04 (cuda 10.0.130)
10.0-runtime-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 6.0.1.5)
10.0-cudnn7-tensorrt7-devel-ubuntu16.04, 10.0-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 7.0.0.11)

Ubuntu 14.04 (No longer updated)

10.0-base-ubuntu14.04 (cuda 10.0.130)
10.0-runtime-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Docker Hub:

[***]

查看更多 cuda 相关镜像 →

rocker/cuda

rocker

集成NVIDIA CUDA库的Rocker镜像，提供R语言环境下的GPU加速计算支持，基于rocker-org/rocker-versioned2项目构建。

英伟达在GitLab仓库（gitlab.com/nvidia/cuda）提供的CUDA（并行计算平台）和cuDNN（深度神经网络加速库）镜像，为开发者提供了预配置的开发环境，支持高效进行并行计算应用开发、深度学习模型训练及推理任务，确保了环境的一致性和部署的便捷性，是构建基于英伟达GPU加速应用的重要资源。

CUDA images to fill gaps and supplement nvidia/cuda

轩辕镜像配置手册

探索更多轩辕镜像的使用方法，找到最适合您系统的配置方式

Docker 配置

登录仓库拉取

通过 Docker 登录认证访问私有仓库

专属域名拉取

无需登录使用专属域名

K8s Containerd

Kubernetes 集群配置 Containerd

K3s

K3s 轻量级 Kubernetes 镜像加速

Dev Containers

VS Code Dev Containers 配置

Podman

Podman 容器引擎配置

Singularity/Apptainer

HPC 科学计算容器配置

其他仓库配置

ghcr、Quay、nvcr 等镜像仓库

系统配置

Linux

在 Linux 系统配置镜像服务

Windows/Mac

在 Docker Desktop 配置镜像

MacOS OrbStack

MacOS OrbStack 容器配置

Docker Compose

Docker Compose 项目配置

NAS 设备

群晖

Synology 群晖 NAS 配置

飞牛

飞牛 fnOS 系统配置镜像

绿联

绿联 NAS 系统配置镜像

威联通

QNAP 威联通 NAS 配置

极空间

极空间 NAS 系统配置服务

网络设备

爱快路由

爱快 iKuai 路由系统配置

宝塔面板

在宝塔面板一键配置镜像

需要其他帮助？请查看我们的常见问题Docker 镜像访问常见问题解答或提交工单

镜像拉取常见问题

使用与功能问题

docker search 报错：专属域名下仅支持 Docker Hub 查询

docker search 报错问题

网页搜不到镜像：Docker Hub 有但轩辕镜像搜索无结果

镜像搜索不到

离线传输镜像：无法直连时用 docker save/load 迁移

离线传输镜像

Docker 插件安装错误：application/vnd.docker.plugin.v1+json

Docker 插件安装错误

WSL 下 Docker 拉取慢：网络与挂载目录影响及优化

WSL 拉取镜像慢

轩辕镜像是否安全？镜像完整性校验（digest）说明

镜像安全性

如何用轩辕镜像拉取镜像？登录方式与专属域名配置

如何拉取镜像

错误码与失败问题

manifest unknown 错误：镜像不存在或标签错误

manifest unknown 错误

TLS/SSL 证书验证失败：Docker pull 时 HTTPS 证书错误

TLS 证书验证失败

DNS 解析超时：无法解析镜像仓库地址或连接超时

DNS 解析超时

410 Gone 错误：Docker 版本过低导致协议不兼容

410 错误：版本过低

402 Payment Required 错误：流量耗尽错误提示

402 错误：流量耗尽

401 UNAUTHORIZED 错误：身份认证失败或登录信息错误

身份认证失败错误

429 Too Many Requests 错误：请求频率超出专业版限制

429 限流错误

Docker login 凭证保存错误：Cannot autolaunch D-Bus（不影响登录）

凭证保存错误

账号 / 计费 / 权限

免费版与专业版区别：功能、限额与使用场景对比

免费版与专业版区别

支持的镜像仓库：Docker Hub、GCR、GHCR、K8s 等列表

轩辕镜像支持的镜像仓库

拉取失败是否扣流量？计费规则说明

拉取失败流量计费

KYSEC 权限不够：麒麟 V10/统信 UOS 下脚本执行被拦截

KYSEC 权限错误

如何申请开具发票？（增值税普票/专票）

开具发票

如何修改网站与仓库登录密码？

修改网站和仓库密码

配置与原理类

registry-mirrors 未生效：仍访问官方仓库或报错的原因

registry-mirrors 未生效

如何去掉镜像名称中的轩辕域名前缀？（docker tag）

去掉域名前缀

如何拉取指定架构镜像？（ARM64/AMD64 等多架构）

拉取指定架构镜像

查看全部问题→

用户好评

来自真实用户的反馈，见证轩辕镜像的优质服务

oldzhang

运维工程师

Linux服务器

"Docker访问体验非常流畅，大镜像也能快速完成下载。"

vistart/cuda

vistart

CUDA CUDNN NCCL TensorRT

17 次收藏下载次数: 0状态：社区镜像维护者：vistart仓库类型：镜像最近更新：4 年前

轩辕镜像，加速的不只是镜像。点击查看

镜像简介版本下载

轩辕镜像，加速的不只是镜像。点击查看

WARNING!!!

All images inherited from ubuntu of Official Version are confusing, PLEASE DO NOT UPDATE! We are now re-build all images from source.

![Build Status]([***]

Introduction

The tag name format is:

registry.cn-hongkong.aliyuncs.com/vistart_public/cuda:<tag name>

After that, if you want to update the image, repeat the above operations.

Latest CUDA 11.0 (Experimental)(CUDA 11.0 for Ubuntu 20.04 is Ready! CUDNN 8.0.2/NCCL 2.7.8/TensorRT 7.1.3 are coming!)

Ubuntu 20.04

11.0-base-ubuntu20.04 (cuda 11.0.3)
11.0-runtime-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8)
11.0-devel-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8)
11.0-cudnn8-runtime-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-devel-ubuntu20.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-tensorrt7-devel-ubuntu20.04, 11.0-ubuntu20.04, latest (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 18.04

11.0-base-ubuntu18.04 (cuda 11.0.3)
11.0-runtime-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8)
11.0-devel-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8)
11.0-cudnn8-runtime-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-devel-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-tensorrt7-devel-ubuntu18.04, 10.2-ubuntu18.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 16.04

11.0-base-ubuntu16.04 (cuda 11.0.3)
11.0-runtime-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8)
11.0-devel-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8)
11.0-cudnn8-runtime-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-devel-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39)
11.0-cudnn8-tensorrt7-devel-ubuntu16.04 (cuda 11.0.3, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

CUDA 10.2

Ubuntu 20.04 (Deprecated)

10.2-base-ubuntu20.04 (cuda 10.2.89)
10.2-runtime-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-runtime-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn7-tensorrt6-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.8)
10.2-cudnn7-tensorrt7-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 7.0.0.11)
10.2-cudnn8-runtime-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39)
10.2-cudnn8-devel-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39)
10.2-cudnn8-tensorrt7-devel-ubuntu20.04, 10.2-ubuntu20.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 18.04

10.2-base-ubuntu18.04 (cuda 10.2.89)
10.2-runtime-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-runtime-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn7-tensorrt6-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.8)
10.2-cudnn7-tensorrt7-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 7.0.0.11)
10.2-cudnn8-runtime-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-devel-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-tensorrt7-devel-ubuntu18.04, 10.2-ubuntu18.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

Ubuntu 16.04

10.2-base-ubuntu16.04 (cuda 10.2.89)
10.2-runtime-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-runtime-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8)
10.2-cudnn7-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn7-tensorrt6-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.8)
10.2-cudnn7-tensorrt7-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 7.0.0.11)
10.2-cudnn8-runtime-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-devel-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 7.6.5.32)
10.2-cudnn8-tensorrt7-devel-ubuntu16.04, 10.2-ubuntu16.04 (cuda 10.2.89, nccl 2.7.8, cudnn 8.0.2.39, tensorrt 7.1.3.4)

CUDA 10.1

Ubuntu 20.04 (Deprecated)

10.1-base-ubuntu20.04 (cuda 10.1.243)
10.1-runtime-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-devel-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn7-tensorrt6-devel-ubuntu20.04, 10.1-ubuntu20.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Ubuntu 18.04

10.1-base-ubuntu18.04 (cuda 10.1.243)
10.1-runtime-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-runtime-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-devel-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-devel-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-cudnn7-tensorrt6-devel-ubuntu18.04, 10.1-ubuntu18.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Ubuntu 16.04

10.1-base-ubuntu16.04 (cuda 10.1.243)
10.1-runtime-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-runtime-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-devel-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn8-devel-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 8.0.2.39)
10.1-cudnn7-tensorrt6-devel-ubuntu16.04, 10.1-ubuntu16.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Ubuntu 14.04

10.1-base-ubuntu14.04 (cuda 10.1.243)
10.1-runtime-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-runtime-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-devel-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8)
10.1-cudnn7-devel-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32)
10.1-cudnn7-tensorrt6-devel-ubuntu14.04 (cuda 10.1.243, nccl 2.7.8, cudnn 7.6.5.32 tensorrt 6.0.1.5)

CUDA 10.0

Ubuntu 20.04 (Deprecated)

10.0-base-ubuntu20.04 (cuda 10.0.130)
10.0-runtime-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32 tensorrt 6.0.1.8)
10.0-cudnn7-tensorrt7-devel-ubuntu20.04, 10.0-ubuntu20.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 7.0.0.11)

Ubuntu 18.04 (No longer updated)

10.0-base-ubuntu18.04 (cuda 10.0.130)
10.0-runtime-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 6.0.1.5)
10.0-cudnn7-tensorrt7-devel-ubuntu18.04, 10.0-ubuntu18.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 7.0.0.11)

Ubuntu 16.04 (No longer updated)

10.0-base-ubuntu16.04 (cuda 10.0.130)
10.0-runtime-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 6.0.1.5)
10.0-cudnn7-tensorrt7-devel-ubuntu16.04, 10.0-ubuntu16.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 7.0.0.11)

Ubuntu 14.04 (No longer updated)

10.0-base-ubuntu14.04 (cuda 10.0.130)
10.0-runtime-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-runtime-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-devel-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4)
10.0-cudnn7-devel-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32)
10.0-cudnn7-tensorrt6-devel-ubuntu14.04 (cuda 10.0.130, nccl 2.6.4, cudnn 7.6.5.32, tensorrt 6.0.1.5)

Docker Hub:

[***]

查看更多 cuda 相关镜像 →

rocker/cuda

rocker

集成NVIDIA CUDA库的Rocker镜像，提供R语言环境下的GPU加速计算支持，基于rocker-org/rocker-versioned2项目构建。

CUDA images to fill gaps and supplement nvidia/cuda

轩辕镜像配置手册

探索更多轩辕镜像的使用方法，找到最适合您系统的配置方式

Docker 配置

登录仓库拉取

通过 Docker 登录认证访问私有仓库

专属域名拉取

无需登录使用专属域名

K8s Containerd

Kubernetes 集群配置 Containerd

K3s

K3s 轻量级 Kubernetes 镜像加速

Dev Containers

VS Code Dev Containers 配置

Podman

Podman 容器引擎配置

Singularity/Apptainer

HPC 科学计算容器配置

其他仓库配置

ghcr、Quay、nvcr 等镜像仓库

系统配置

Linux

在 Linux 系统配置镜像服务

Windows/Mac

在 Docker Desktop 配置镜像

MacOS OrbStack

MacOS OrbStack 容器配置

Docker Compose

Docker Compose 项目配置

NAS 设备

群晖

Synology 群晖 NAS 配置

飞牛

飞牛 fnOS 系统配置镜像

绿联

绿联 NAS 系统配置镜像

威联通

QNAP 威联通 NAS 配置

极空间

极空间 NAS 系统配置服务

网络设备

爱快路由

爱快 iKuai 路由系统配置

宝塔面板

在宝塔面板一键配置镜像

需要其他帮助？请查看我们的常见问题Docker 镜像访问常见问题解答或提交工单

镜像拉取常见问题

使用与功能问题

docker search 报错：专属域名下仅支持 Docker Hub 查询

docker search 报错问题

网页搜不到镜像：Docker Hub 有但轩辕镜像搜索无结果

镜像搜索不到

离线传输镜像：无法直连时用 docker save/load 迁移

离线传输镜像

Docker 插件安装错误：application/vnd.docker.plugin.v1+json

Docker 插件安装错误

WSL 下 Docker 拉取慢：网络与挂载目录影响及优化

WSL 拉取镜像慢

轩辕镜像是否安全？镜像完整性校验（digest）说明

镜像安全性

如何用轩辕镜像拉取镜像？登录方式与专属域名配置

如何拉取镜像

错误码与失败问题

manifest unknown 错误：镜像不存在或标签错误

manifest unknown 错误

TLS/SSL 证书验证失败：Docker pull 时 HTTPS 证书错误

TLS 证书验证失败

DNS 解析超时：无法解析镜像仓库地址或连接超时

DNS 解析超时

410 Gone 错误：Docker 版本过低导致协议不兼容

410 错误：版本过低

402 Payment Required 错误：流量耗尽错误提示

402 错误：流量耗尽

401 UNAUTHORIZED 错误：身份认证失败或登录信息错误

身份认证失败错误

429 Too Many Requests 错误：请求频率超出专业版限制

429 限流错误

Docker login 凭证保存错误：Cannot autolaunch D-Bus（不影响登录）

凭证保存错误

账号 / 计费 / 权限

免费版与专业版区别：功能、限额与使用场景对比

免费版与专业版区别

支持的镜像仓库：Docker Hub、GCR、GHCR、K8s 等列表

轩辕镜像支持的镜像仓库

拉取失败是否扣流量？计费规则说明

拉取失败流量计费

KYSEC 权限不够：麒麟 V10/统信 UOS 下脚本执行被拦截

KYSEC 权限错误

如何申请开具发票？（增值税普票/专票）

开具发票

如何修改网站与仓库登录密码？

修改网站和仓库密码

配置与原理类

registry-mirrors 未生效：仍访问官方仓库或报错的原因

registry-mirrors 未生效

如何去掉镜像名称中的轩辕域名前缀？（docker tag）

去掉域名前缀

如何拉取指定架构镜像？（ARM64/AMD64 等多架构）

拉取指定架构镜像

查看全部问题→

用户好评

来自真实用户的反馈，见证轩辕镜像的优质服务

oldzhang

运维工程师

Linux服务器

"Docker访问体验非常流畅，大镜像也能快速完成下载。"

vistart/cuda Docker 镜像 - 轩辕镜像 | Docker 镜像高效稳定拉取服务

WARNING!!!

Introduction

Latest CUDA 11.0 (Experimental)(CUDA 11.0 for Ubuntu 20.04 is Ready! CUDNN 8.0.2/NCCL 2.7.8/TensorRT 7.1.3 are coming!)

Ubuntu 20.04

Ubuntu 18.04

Ubuntu 16.04

CUDA 10.2

Ubuntu 20.04 (Deprecated)

Ubuntu 18.04

Ubuntu 16.04

CUDA 10.1

Ubuntu 20.04 (Deprecated)

Ubuntu 18.04

Ubuntu 16.04

Ubuntu 14.04

CUDA 10.0

Ubuntu 20.04 (Deprecated)

Ubuntu 18.04 (No longer updated)

Ubuntu 16.04 (No longer updated)

Ubuntu 14.04 (No longer updated)

Docker Hub:

轩辕镜像配置手册

Docker 配置

登录仓库拉取

专属域名拉取

K8s Containerd

K3s

Dev Containers

Podman

Singularity/Apptainer

其他仓库配置

系统配置

Linux

Windows/Mac

MacOS OrbStack

Docker Compose

NAS 设备

群晖

飞牛

绿联

威联通

极空间

网络设备

爱快路由

宝塔面板

镜像拉取常见问题

使用与功能问题

docker search 报错：专属域名下仅支持 Docker Hub 查询

网页搜不到镜像：Docker Hub 有但轩辕镜像搜索无结果

离线传输镜像：无法直连时用 docker save/load 迁移

Docker 插件安装错误：application/vnd.docker.plugin.v1+json

WSL 下 Docker 拉取慢：网络与挂载目录影响及优化

轩辕镜像是否安全？镜像完整性校验（digest）说明

如何用轩辕镜像拉取镜像？登录方式与专属域名配置

错误码与失败问题

manifest unknown 错误：镜像不存在或标签错误

TLS/SSL 证书验证失败：Docker pull 时 HTTPS 证书错误

DNS 解析超时：无法解析镜像仓库地址或连接超时

410 Gone 错误：Docker 版本过低导致协议不兼容

402 Payment Required 错误：流量耗尽错误提示

401 UNAUTHORIZED 错误：身份认证失败或登录信息错误

429 Too Many Requests 错误：请求频率超出专业版限制

Docker login 凭证保存错误：Cannot autolaunch D-Bus（不影响登录）

账号 / 计费 / 权限

免费版与专业版区别：功能、限额与使用场景对比

支持的镜像仓库：Docker Hub、GCR、GHCR、K8s 等列表

拉取失败是否扣流量？计费规则说明

KYSEC 权限不够：麒麟 V10/统信 UOS 下脚本执行被拦截

如何申请开具发票？（增值税普票/专票）

如何修改网站与仓库登录密码？

配置与原理类

registry-mirrors 未生效：仍访问官方仓库或报错的原因

如何去掉镜像名称中的轩辕域名前缀？（docker tag）

如何拉取指定架构镜像？（ARM64/AMD64 等多架构）

用户好评

vistart/cuda Docker 镜像 - 轩辕镜像 | Docker 镜像高效稳定拉取服务

WARNING!!!

Introduction

Latest CUDA 11.0 (Experimental)(CUDA 11.0 for Ubuntu 20.04 is Ready! CUDNN 8.0.2/NCCL 2.7.8/TensorRT 7.1.3 are coming!)