在Ubuntu 16.04上製作 NVIDIA cuDNN Docker image
第七步：使用這些Docker image本文講述了如何在Kde Ubuntu 16.04.03 (LTS發行版）上製作（編譯）NVIDIA cuDNN Docker image文件。
第一步：製作好NVIDIA CUDA Docker image：
參考Gemfield專欄文章：在Ubuntu 16.04上製作 NVIDIA CUDA Docker image
第二步：build NVIDIA cuDNN 運行時的 Docker image：
1，取得Ubuntu版的NVIDIA cuDNN 6.0 runtime版本的Dockerfile：
下載：8.0/runtime/cudnn6/Dockerfile · ubuntu16.04 · nvidia / cuda
然後放置到當前目錄（本步驟沒有使用上面的官方版本，而是使用了下面的版本），這個Dockerfile使用了 「gemfield/ubuntu-runtime-cuda」 作為父 image。
FROM gemfield/ubuntu-runtime-cuda:8.0LABEL maintainer "Gemfield <gemfield@civilnet.cn>"RUN echo "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 /" > /etc/apt/sources.list.d/nvidia-ml.listENV CUDNN_VERSION 6.0.21LABEL com.nvidia.cudnn.version="${CUDNN_VERSION}"RUN apt-get update && apt-get install -y --no-install-recommends libcudnn6=$CUDNN_VERSION-1+cuda8.0 && rm -rf /var/lib/apt/lists/*
2，在包含上面Dockerfile的當前目錄，執行Docker build命令來生成"gemfield/ubuntu-runtime-cudnn:6.0" Docker image：
gemfield@ai:~/AI/dockerfiles/cudnn/runtime$ docker build -t "gemfield/ubuntu-runtime-cudnn:6.0" . Sending build context to Docker daemon 2.048kBStep 1/6 : FROM gemfield/ubuntu-runtime-cuda:8.0 ---> d0b82ef8f4d2Step 2/6 : LABEL maintainer "Gemfield <gemfield@civilnet.cn>" ---> Using cache ---> c103d5ca0de3Step 3/6 : RUN echo "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 /" > /etc/apt/sources.list.d/nvidia-ml.list ---> Running in 34d93da88528 ---> d8c2c79a024bRemoving intermediate container 34d93da88528Step 4/6 : ENV CUDNN_VERSION 6.0.21 ---> Running in e16c8fb33d93 ---> 27617abd2787Removing intermediate container e16c8fb33d93Step 5/6 : LABEL com.nvidia.cudnn.version "${CUDNN_VERSION}" ---> Running in 9cddec34fd9f ---> 07a1af4534e4Removing intermediate container 9cddec34fd9fStep 6/6 : RUN apt-get update && apt-get install -y --no-install-recommends libcudnn6=$CUDNN_VERSION-1+cuda8.0 && rm -rf /var/lib/apt/lists/* ---> Running in eaa0a052e3f8Get:1 http://security.ubuntu.com/ubuntu xenial-security InRelease [102 kB]Ign:2 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 InReleaseGet:3 http://archive.ubuntu.com/ubuntu xenial InRelease [247 kB]Ign:4 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 InReleaseGet:5 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 Release [564 B]Get:6 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 Release [564 B]Get:7 http://security.ubuntu.com/ubuntu xenial-security/universe Sources [46.8 kB]Get:8 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 Release.gpg [801 B]Get:9 http://security.ubuntu.com/ubuntu xenial-security/main amd64 Packages [440 kB]Get:10 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 Release.gpg [801 B]Get:11 http://security.ubuntu.com/ubuntu xenial-security/restricted amd64 Packages [12.8 kB]Get:12 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 Packages [66.0 kB]Get:13 http://security.ubuntu.com/ubuntu xenial-security/universe amd64 Packages [204 kB]Get:14 http://security.ubuntu.com/ubuntu xenial-security/multiverse amd64 Packages [2935 B]Get:15 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 Packages [8408 B]Get:16 http://archive.ubuntu.com/ubuntu xenial-updates InRelease [102 kB]Get:17 http://archive.ubuntu.com/ubuntu xenial-backports InRelease [102 kB]Get:18 http://archive.ubuntu.com/ubuntu xenial/universe Sources [9802 kB]Get:19 http://archive.ubuntu.com/ubuntu xenial/main amd64 Packages [1558 kB]Get:20 http://archive.ubuntu.com/ubuntu xenial/restricted amd64 Packages [14.1 kB]Get:21 http://archive.ubuntu.com/ubuntu xenial/universe amd64 Packages [9827 kB]Get:22 http://archive.ubuntu.com/ubuntu xenial/multiverse amd64 Packages [176 kB]Get:23 http://archive.ubuntu.com/ubuntu xenial-updates/universe Sources [214 kB]Get:24 http://archive.ubuntu.com/ubuntu xenial-updates/main amd64 Packages [799 kB]Get:25 http://archive.ubuntu.com/ubuntu xenial-updates/restricted amd64 Packages [13.6 kB]Get:26 http://archive.ubuntu.com/ubuntu xenial-updates/universe amd64 Packages [666 kB]Get:27 http://archive.ubuntu.com/ubuntu xenial-updates/multiverse amd64 Packages [17.5 kB]Get:28 http://archive.ubuntu.com/ubuntu xenial-backports/main amd64 Packages [5177 B]Get:29 http://archive.ubuntu.com/ubuntu xenial-backports/universe amd64 Packages [6236 B]Fetched 24.4 MB in 49s (497 kB/s)Reading package lists...Reading package lists...Building dependency tree...Reading state information...The following NEW packages will be installed: libcudnn60 upgraded, 1 newly installed, 0 to remove and 0 not upgraded.Need to get 68.5 MB of archives.After this operation, 154 MB of additional disk space will be used.Get:1 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 libcudnn6 6.0.21-1+cuda8.0 [68.5 MB]debconf: delaying package configuration, since apt-utils is not installedFetched 68.5 MB in 1min 44s (658 kB/s)Selecting previously unselected package libcudnn6.(Reading database ... 4887 files and directories currently installed.)Preparing to unpack .../libcudnn6_6.0.21-1+cuda8.0_amd64.deb ...Unpacking libcudnn6 (6.0.21-1+cuda8.0) ...Processing triggers for libc-bin (2.23-0ubuntu9) ...Setting up libcudnn6 (6.0.21-1+cuda8.0) ...Processing triggers for libc-bin (2.23-0ubuntu9) ... ---> ccfac2fb896cRemoving intermediate container eaa0a052e3f8Successfully built ccfac2fb896cSuccessfully tagged gemfield/ubuntu-runtime-cudnn:6.0gemfield@ai:~/AI/dockerfiles/cudnn/runtime$
3，運行docker images命令來查看生成的image：
gemfield@ai:~/AI/dockerfiles/cudnn/runtime$ docker imagesREPOSITORY TAG IMAGE ID CREATED SIZEgemfield/ubuntu-runtime-cudnn 6.0 ccfac2fb896c 5 minutes ago 935MBgemfield/ubuntu-devel-cuda 8.0 ab0c5b19d347 About an hour ago 1.68GBgemfield/ubuntu-runtime-cuda 8.0 d0b82ef8f4d2 2 hours ago 780MBnvidia/cuda latest f1193af0e70e 2 weeks ago 1.68GBubuntu 16.04 ccc7a11d65b1 2 weeks ago 120MBhello-world latest 1815c82652c0 2 months ago 1.84kBgemfield@ai:~/AI/dockerfiles/cudnn/runtime$
第三步：build NVIDIA cuDNN 開發版的 Docker image：
1，取得Ubuntu版的NVIDIA cuDNN 6.0 devel版本的Dockerfile：
下載：8.0/devel/cudnn6/Dockerfile · ubuntu16.04 · nvidia / cuda
然後放置到當前目錄（本步驟沒有使用上面的官方版本，而是使用了下面的版本），這個Dockerfile使用了 「gemfield/ubuntu-devel-cuda」 作為父 image。
FROM gemfield/ubuntu-devel-cuda:8.0LABEL maintainer "Gemfield <gemfield@civilnet.cn>"RUN echo "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 /" > /etc/apt/sources.list.d/nvidia-ml.listENV CUDNN_VERSION 6.0.21LABEL com.nvidia.cudnn.version="${CUDNN_VERSION}"RUN apt-get update && apt-get install -y --no-install-recommends libcudnn6=$CUDNN_VERSION-1+cuda8.0 libcudnn6-dev=$CUDNN_VERSION-1+cuda8.0 && rm -rf /var/lib/apt/lists/*
2，在包含上面Dockerfile的當前目錄，執行Docker build命令來生成"gemfield/ubuntu-devel-cudnn:6.0" Docker image：
gemfield@ai:~/AI/dockerfiles/cudnn/devel$ docker build -t "gemfield/ubuntu-devel-cudnn:6.0" .Sending build context to Docker daemon 2.56kBStep 1/6 : FROM gemfield/ubuntu-devel-cuda:8.0 ---> ab0c5b19d347Step 2/6 : LABEL maintainer "Gemfield <gemfield@civilnet.cn>" ---> Running in d8ed897a46ee ---> 741522f70e05Removing intermediate container d8ed897a46eeStep 3/6 : RUN echo "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 /" > /etc/apt/sources.list.d/nvidia-ml.list ---> Running in 7ed8328e5e17 ---> 60273b49a65aRemoving intermediate container 7ed8328e5e17Step 4/6 : ENV CUDNN_VERSION 6.0.21 ---> Running in 33040bd35d03 ---> 4832cc2661bcRemoving intermediate container 33040bd35d03Step 5/6 : LABEL com.nvidia.cudnn.version "${CUDNN_VERSION}" ---> Running in bf981254f3b6 ---> 83a256035a47Removing intermediate container bf981254f3b6Step 6/6 : RUN apt-get update && apt-get install -y --no-install-recommends libcudnn6=$CUDNN_VERSION-1+cuda8.0 libcudnn6-dev=$CUDNN_VERSION-1+cuda8.0 && rm -rf /var/lib/apt/lists/* ---> Running in 7927fb83c83bGet:1 http://security.ubuntu.com/ubuntu xenial-security InRelease [102 kB]Get:2 http://archive.ubuntu.com/ubuntu xenial InRelease [247 kB]Ign:3 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 InReleaseIgn:4 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 InReleaseGet:5 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 Release [564 B]Get:6 http://security.ubuntu.com/ubuntu xenial-security/universe Sources [46.8 kB]Get:7 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 Release [564 B]Get:8 http://security.ubuntu.com/ubuntu xenial-security/main amd64 Packages [441 kB]Get:9 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 Release.gpg [801 B]Get:10 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 Release.gpg [801 B]Get:11 http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64 Packages [66.0 kB]Get:12 http://security.ubuntu.com/ubuntu xenial-security/restricted amd64 Packages [12.8 kB]Get:13 http://security.ubuntu.com/ubuntu xenial-security/universe amd64 Packages [204 kB]Get:14 http://security.ubuntu.com/ubuntu xenial-security/multiverse amd64 Packages [2935 B]Get:15 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 Packages [8408 B]Get:16 http://archive.ubuntu.com/ubuntu xenial-updates InRelease [102 kB]Get:17 http://archive.ubuntu.com/ubuntu xenial-backports InRelease [102 kB]Get:18 http://archive.ubuntu.com/ubuntu xenial/universe Sources [9802 kB]Get:19 http://archive.ubuntu.com/ubuntu xenial/main amd64 Packages [1558 kB]Get:20 http://archive.ubuntu.com/ubuntu xenial/restricted amd64 Packages [14.1 kB]Get:21 http://archive.ubuntu.com/ubuntu xenial/universe amd64 Packages [9827 kB]Get:22 http://archive.ubuntu.com/ubuntu xenial/multiverse amd64 Packages [176 kB]Get:23 http://archive.ubuntu.com/ubuntu xenial-updates/universe Sources [214 kB]Get:24 http://archive.ubuntu.com/ubuntu xenial-updates/main amd64 Packages [799 kB]Get:25 http://archive.ubuntu.com/ubuntu xenial-updates/restricted amd64 Packages [13.6 kB]Get:26 http://archive.ubuntu.com/ubuntu xenial-updates/universe amd64 Packages [666 kB]Get:27 http://archive.ubuntu.com/ubuntu xenial-updates/multiverse amd64 Packages [17.5 kB]Get:28 http://archive.ubuntu.com/ubuntu xenial-backports/main amd64 Packages [5177 B]Get:29 http://archive.ubuntu.com/ubuntu xenial-backports/universe amd64 Packages [6236 B]Fetched 24.4 MB in 34s (708 kB/s)Reading package lists...Reading package lists...Building dependency tree...Reading state information...The following NEW packages will be installed: libcudnn6 libcudnn6-dev0 upgraded, 2 newly installed, 0 to remove and 0 not upgraded.Need to get 128 MB of archives.After this operation, 298 MB of additional disk space will be used.Get:1 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 libcudnn6 6.0.21-1+cuda8.0 [68.5 MB]Get:2 http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64 libcudnn6-dev 6.0.21-1+cuda8.0 [59.9 MB]debconf: delaying package configuration, since apt-utils is not installedFetched 128 MB in 2min 44s (779 kB/s)Selecting previously unselected package libcudnn6.(Reading database ... 11141 files and directories currently installed.)Preparing to unpack .../libcudnn6_6.0.21-1+cuda8.0_amd64.deb ...Unpacking libcudnn6 (6.0.21-1+cuda8.0) ...Selecting previously unselected package libcudnn6-dev.Preparing to unpack .../libcudnn6-dev_6.0.21-1+cuda8.0_amd64.deb ...Unpacking libcudnn6-dev (6.0.21-1+cuda8.0) ...Processing triggers for libc-bin (2.23-0ubuntu9) ...Setting up libcudnn6 (6.0.21-1+cuda8.0) ...Setting up libcudnn6-dev (6.0.21-1+cuda8.0) ...update-alternatives: using /usr/include/x86_64-linux-gnu/cudnn_v6.h to provide /usr/include/cudnn.h (libcudnn) in auto modeProcessing triggers for libc-bin (2.23-0ubuntu9) ... ---> 441e2eab5c7eRemoving intermediate container 7927fb83c83bSuccessfully built 441e2eab5c7eSuccessfully tagged gemfield/ubuntu-devel-cudnn:6.0gemfield@ai:~/AI/dockerfiles/cudnn/devel$
3，運行docker images命令來查看生成的image：
gemfield@ai:~/AI/dockerfiles/cudnn/devel$ docker imagesREPOSITORY TAG IMAGE ID CREATED SIZEgemfield/ubuntu-devel-cudnn 6.0 441e2eab5c7e 56 seconds ago 1.98GBgemfield/ubuntu-runtime-cudnn 6.0 ccfac2fb896c 16 hours ago 935MBgemfield/ubuntu-devel-cuda 8.0 ab0c5b19d347 17 hours ago 1.68GBgemfield/ubuntu-runtime-cuda 8.0 d0b82ef8f4d2 18 hours ago 780MBnvidia/cuda latest f1193af0e70e 2 weeks ago 1.68GBubuntu 16.04 ccc7a11d65b1 2 weeks ago 120MBhello-world latest 1815c82652c0 2 months ago 1.84kBgemfield@ai:~/AI/dockerfiles/cudnn/devel$
第四步：上傳這兩個cuDNN Docker image到Docker Hub：
參考Gemfield專欄文章：在Ubuntu 16.04上製作 NVIDIA CUDA Docker image
第五步：使用這些Docker image
參考Gemfield專欄文章：在Ubuntu 16.04上製作 NVIDIA CUDA Docker image
推薦閱讀：
TAG:NVIDIA英偉達 | Docker | Caffe深度學習框架 |