/
Dockerfile
256 lines (225 loc) · 10 KB
/
Dockerfile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
# Tag: nvidia/cuda:9.0-cudnn7-devel-ubuntu16.04
# Created: 2017-11-21T06:34:14.675603521Z
# Label: com.nvidia.build.id: 41212533
# Label: com.nvidia.build.ref: e0edb5359ecb7bd3d86f0c9bfa18c2260b741ebb
# Label: com.nvidia.cuda.version: 9.0.176
# Label: com.nvidia.cudnn.version: 7.0.4.31
# Label: com.nvidia.nccl.version: 2.1.2
#
# To build, run from the parent with the command line:
# docker build -t <image name> -f CNTK-GPU-Image/Dockerfile .
# Ubuntu 16.04.5
FROM nvidia/cuda@sha256:33add9c50ab76b8f3a92187c0418ed600d5bea27690fda40711122fdc28ce2f4
RUN apt-get update && apt-get install -y --no-install-recommends \
autotools-dev \
build-essential \
git \
gfortran-multilib \
libavcodec-dev \
libavformat-dev \
libjasper-dev \
libjpeg-dev \
libpng-dev \
liblapacke-dev \
libswscale-dev \
libtiff-dev \
pkg-config \
wget \
zlib1g-dev \
# OPENSSL support
libssl-dev \
openssl \
# Protobuf
ca-certificates \
curl \
libcurl4-openssl-dev \
unzip \
# For Kaldi
python-dev \
automake \
libtool-bin \
autoconf \
subversion \
# For Kaldi's dependencies
libapr1 libaprutil1 libltdl-dev libltdl7 libserf-1-1 libsigsegv2 libsvn1 m4 \
# For Java Bindings
openjdk-8-jdk \
# For SWIG
libpcre3-dev \
# For graphics managed lib
libgdiplus \
# .NET Core SDK
apt-transport-https && \
# Cleanup
rm -rf /var/lib/apt/lists/*
ARG CMAKE_DOWNLOAD_VERSION=3.11
ARG CMAKE_BUILD_VERSION=4
RUN DEBIAN_FRONTEND=noninteractive && \
wget --no-verbose https://cmake.org/files/v$CMAKE_DOWNLOAD_VERSION/cmake-$CMAKE_DOWNLOAD_VERSION.$CMAKE_BUILD_VERSION.tar.gz && \
tar -xzvf cmake-$CMAKE_DOWNLOAD_VERSION.$CMAKE_BUILD_VERSION.tar.gz && \
cd cmake-$CMAKE_DOWNLOAD_VERSION.$CMAKE_BUILD_VERSION && \
./bootstrap --system-curl -- -DCMAKE_USE_OPENSSL=ON && \
make -j $(nproc) install && \
cd .. && \
rm -rf cmake-$CMAKE_DOWNLOAD_VERSION.$CMAKE_BUILD_VERSION
RUN OPENMPI_VERSION=1.10.3 && \
wget -q -O - https://www.open-mpi.org/software/ompi/v1.10/downloads/openmpi-${OPENMPI_VERSION}.tar.gz | tar -xzf - && \
cd openmpi-${OPENMPI_VERSION} && \
./configure --prefix=/usr/local/mpi && \
make -j"$(nproc)" install && \
rm -rf /openmpi-${OPENMPI_VERSION}
ENV PATH /usr/local/mpi/bin:$PATH
ENV LD_LIBRARY_PATH /usr/local/mpi/lib:$LD_LIBRARY_PATH
RUN LIBZIP_VERSION=1.1.2 && \
wget -q -O - https://libzip.org/download/libzip-${LIBZIP_VERSION}.tar.gz | tar -xzf - && \
cd libzip-${LIBZIP_VERSION} && \
./configure && \
make -j"$(nproc)" install && \
rm -rf /libzip-${LIBZIP_VERSION}
ENV LD_LIBRARY_PATH /usr/local/lib:$LD_LIBRARY_PATH
RUN wget -q -O - https://github.com/NVlabs/cub/archive/1.7.4.tar.gz | tar -C /usr/local -xzf -
RUN OPENCV_VERSION=3.1.0 && \
wget -q -O - https://github.com/Itseez/opencv/archive/${OPENCV_VERSION}.tar.gz | tar -xzf - && \
cd opencv-${OPENCV_VERSION} && \
cmake -DWITH_CUDA=OFF -DCMAKE_BUILD_TYPE=RELEASE -DCMAKE_INSTALL_PREFIX=/usr/local/opencv-${OPENCV_VERSION} . && \
make -j"$(nproc)" install && \
rm -rf /opencv-${OPENCV_VERSION}
RUN OPENBLAS_VERSION=0.2.18 && \
wget -q -O - https://github.com/xianyi/OpenBLAS/archive/v${OPENBLAS_VERSION}.tar.gz | tar -xzf - && \
cd OpenBLAS-${OPENBLAS_VERSION} && \
make -j"$(nproc)" USE_OPENMP=1 | tee make.log && \
grep -qF 'OpenBLAS build complete. (BLAS CBLAS LAPACK LAPACKE)' make.log && \
grep -qF 'Use OpenMP in the multithreading.' make.log && \
make PREFIX=/usr/local/openblas install && \
rm -rf /OpenBLAS-${OPENBLAS_VERSION}
ENV LD_LIBRARY_PATH /usr/local/openblas/lib:$LD_LIBRARY_PATH
# Install Boost
RUN BOOST_VERSION=1_60_0 && \
BOOST_DOTTED_VERSION=$(echo $BOOST_VERSION | tr _ .) && \
wget -q -O - https://sourceforge.net/projects/boost/files/boost/${BOOST_DOTTED_VERSION}/boost_${BOOST_VERSION}.tar.gz/download | tar -xzf - && \
cd boost_${BOOST_VERSION} && \
./bootstrap.sh --prefix=/usr/local/boost-${BOOST_DOTTED_VERSION} --with-libraries=filesystem,system,test && \
./b2 -d0 -j"$(nproc)" install && \
rm -rf /boost_${BOOST_VERSION}
# Install Protobuf
RUN PROTOBUF_VERSION=3.1.0 \
PROTOBUF_STRING=protobuf-$PROTOBUF_VERSION && \
wget -O - --no-verbose https://github.com/google/protobuf/archive/v${PROTOBUF_VERSION}.tar.gz | tar -xzf - && \
cd $PROTOBUF_STRING && \
./autogen.sh && \
./configure CFLAGS=-fPIC CXXFLAGS=-fPIC --disable-shared --prefix=/usr/local/$PROTOBUF_STRING && \
make -j $(nproc) install && \
cd .. && \
rm -rf $PROTOBUF_STRING
# Install MKLDNN and MKLML
ARG MKLDNN_VERSION=0.14
ARG MKLDNN_LONG_VERSION=mklml_lnx_2018.0.3.20180406
RUN mkdir /usr/local/mklml && \
wget --no-verbose -O - https://github.com/01org/mkl-dnn/releases/download/v${MKLDNN_VERSION}/${MKLDNN_LONG_VERSION}.tgz | \
tar -xzf - -C /usr/local/mklml && \
MKLDNN_STRING=mkl-dnn-${MKLDNN_VERSION} && \
wget --no-verbose -O - https://github.com/01org/mkl-dnn/archive/v${MKLDNN_VERSION}.tar.gz | tar -xzf - && \
cd ${MKLDNN_STRING} && \
ln -s /usr/local external && \
mkdir -p build && \
cd build && \
cmake .. -DCMAKE_INSTALL_PREFIX=/ && \
make && \
make install DESTDIR=/usr/local && \
make install DESTDIR=/usr/local/mklml/${MKLDNN_LONG_VERSION} && \
cd ../.. && \
rm -rf ${MKLDNN_STRING}
# Install Kaldi
ENV KALDI_VERSION=c024e8aa
ENV KALDI_PATH /usr/local/kaldi-$KALDI_VERSION
RUN mv /bin/sh /bin/sh.orig && \
ln -s -f /bin/bash /bin/sh && \
mkdir $KALDI_PATH && \
wget --no-verbose -O - https://github.com/kaldi-asr/kaldi/archive/$KALDI_VERSION.tar.gz | tar -xzf - --strip-components=1 -C $KALDI_PATH && \
cd $KALDI_PATH && \
cd tools && \
perl -pi -e 's/^# (OPENFST_VERSION = 1.4.1)$/\1/' Makefile && \
./extras/check_dependencies.sh && \
make -j $(nproc) all && \
cd ../src && \
# remove Fermi support as CUDA 9 no longer works on it
perl -pi -e 's/-gencode arch=compute_20,code=sm_20//' cudamatrix/Makefile && \
./configure --openblas-root=/usr/local/openblas --shared && \
make -j $(nproc) depend && \
make -j $(nproc) all && \
# Remove some unneeded stuff in $KALDI_PATH to reduce size
find $KALDI_PATH -name '*.o' -print0 | xargs -0 rm && \
for dir in $KALDI_PATH/src/*bin; do make -C $dir clean; done && \
mv -f /bin/sh.orig /bin/sh
## PYTHON
# Commit that will be used for Python environment creation (and later, compilation)
ARG COMMIT=master
# Swig
RUN cd /root && \
wget -q http://prdownloads.sourceforge.net/swig/swig-3.0.10.tar.gz -O - | tar xvfz - && \
cd swig-3.0.10 && \
# Note: we specify --without-alllang to suppress building tests and examples for specific languages.
./configure --without-alllang && \
make -j $(nproc) && \
make install
COPY ./Patches /tmp/patches
RUN /tmp/patches/patch_swig.sh /usr/local/share/swig/3.0.10 && \
rm -rfd /tmp/patches
# .NET Core SDK
RUN cd /tmp && \
(wget -q packages-microsoft-prod.deb https://packages.microsoft.com/config/ubuntu/16.04/packages-microsoft-prod.deb || true) && \
dpkg -i packages-microsoft-prod.deb && \
apt-get update && \
apt-get --yes install dotnet-sdk-2.1.200 && \
rm /tmp/packages-microsoft-prod.deb && \
apt-get --yes autoremove && \
rm -rf /var/lib/apt/lists/*
# Anaconda
RUN wget -q https://repo.continuum.io/archive/Anaconda3-4.2.0-Linux-x86_64.sh && \
bash Anaconda3-4.2.0-Linux-x86_64.sh -b && \
rm Anaconda3-4.2.0-Linux-x86_64.sh
RUN CONDA_ENV_PATH=/tmp/conda-linux-cntk-py35-environment.yml; \
wget -q https://raw.githubusercontent.com/Microsoft/CNTK/$COMMIT/Scripts/install/linux/conda-linux-cntk-py35-environment.yml -O "$CONDA_ENV_PATH" && \
/root/anaconda3/bin/conda env create -p /root/anaconda3/envs/cntk-py35 --file "$CONDA_ENV_PATH" && \
rm -f "$CONDA_ENV_PATH"
ENV PATH /root/anaconda3/envs/cntk-py35/bin:$PATH
WORKDIR /cntk
# Allow CNTK's configure to pick up GDK and CuDNN where it expects it.
# (Note: $CUDNN_VERSION is defined by NVidia's base image)
# TODO hack, CNTK configure should be improved.
RUN _CUDNN_VERSION=$(echo $CUDNN_VERSION | cut -d. -f1-2) && \
mkdir -p /usr/local/cudnn-$_CUDNN_VERSION/cuda/include && \
ln -s /usr/include/cudnn.h /usr/local/cudnn-$_CUDNN_VERSION/cuda/include/cudnn.h && \
mkdir -p /usr/local/cudnn-$_CUDNN_VERSION/cuda/lib64 && \
ln -s /etc/alternatives/libcudnn_so /usr/local/cudnn-$_CUDNN_VERSION/cuda/lib64/libcudnn.so && \
ln -s /usr/local/cudnn-$_CUDNN_VERSION && \
mkdir -p /usr/src/gdk/nvml/lib && \
cp -av /usr/local/cuda/lib64/stubs/libnvidia-ml* /usr/src/gdk/nvml/lib && \
cp -av /usr/local/cuda/lib64/stubs/libnvidia-ml.so /usr/src/gdk/nvml/lib/libnvidia-ml.so.1 && \
mkdir -p /usr/include/nvidia/gdk && \
cp -av /usr/local/cuda/include/nvml.h /usr/include/nvidia/gdk/nvml.h
# Build CNTK
RUN git clone --depth=1 --recursive -b $COMMIT https://github.com/Microsoft/CNTK.git cntksrc && \
cd cntksrc && \
MKLML_VERSION_DETAIL=${MKLDNN_LONG_VERSION} && \
CONFIGURE_OPTS="\
--with-kaldi=${KALDI_PATH} \
--with-py35-path=/root/anaconda3/envs/cntk-py35" && \
mkdir -p build/gpu/release && \
cd build/gpu/release && \
../../../configure $CONFIGURE_OPTS --with-openblas=/usr/local/openblas && \
make -j"$(nproc)" all && \
cd ../../.. && \
mkdir -p build-mkl/gpu/release && \
cd build-mkl/gpu/release && \
../../../configure $CONFIGURE_OPTS --with-mkl=/usr/local/mklml/$MKLML_VERSION_DETAIL && \
make -j"$(nproc)" all
RUN cd cntksrc/Examples/Image/DataSets/CIFAR-10 && \
python install_cifar10.py && \
cd ../../../..
RUN cd cntksrc/Examples/Image/DataSets/MNIST && \
python install_mnist.py && \
cd ../../../..
ENV PATH=/cntk/cntksrc/build/gpu/release/bin:$PATH PYTHONPATH=/cntk/cntksrc/bindings/python LD_LIBRARY_PATH=/cntk/cntksrc/bindings/python/cntk/libs:$LD_LIBRARY_PATH
# Install CNTK as the default backend for Keras
ENV KERAS_BACKEND=cntk