NGC PyTorch 24.12 Dockerfile & requirements.txt

vendor/ngc-pytorch/Dockerfile.24.12-pytorch2.6-py312-cuda12.6

@@ -0,0 +1,312 @@
+FROM nvcr.io/nvidia/pytorch:24.12-py3
+# NVIDIA NGC PyTorch with Python 3.12
+
+ENV CODESERVER=4.96.4
+ENV DEBIAN_FRONTEND=noninteractive \
+    MPLBACKEND=Svg \
+    PIP_IGNORE_INSTALLED=0 \
+    PYTHONUNBUFFERED=1 \
+    LD_LIBRARY_PATH="/usr/local/lib/python3.12/dist-packages/torch/lib:/usr/local/lib/python3.12/dist-packages/torch_tensorrt/lib:/usr/local/cuda/compat/lib:/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/extras/CUPTI/lib64:/usr/local/cuda-12.6:/usr/local/cuda-12.6/include:/usr/include/x86_64-linux-gnu:/opt/hpcx/ucc/lib" \
+    PATH="/usr/local/lib/python3.12/dist-packages/torch_tensorrt/bin:/usr/local/mpi/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/local/ucx/bin:/opt/tensorrt/bin" \
+    BNB_CUDA_VERSION=126 \
+    CPLUS_INCLUDE_PATH=/usr/include/gdal \
+    C_INCLUDE_PATH=/usr/include/gdal \
+    LANG=C.UTF-8
+
+#RUN apt remove libappstream3
+#
+RUN apt-key adv --refresh-keys --keyserver keyserver.ubuntu.com 
+RUN apt-get update
+RUN apt-get install -y --no-install-recommends \
+	automake \
+	bison \
+	build-essential \
+	bzip2 \
+	ca-certificates \
+	cabextract \
+	chrpath \
+	devscripts \
+	dh-make \
+	dirmngr \
+	dkms \
+#	dpatch \
+	ethtool \
+	fakeroot \
+	ffmpeg \
+	flex \
+	fonts-nanum \
+	fonts-nanum-coding \
+	fonts-nanum-extra \
+	g++ \
+	gcc \
+	gdal-bin \
+	gfortran \
+	gir1.2-gtk-3.0 \
+	git \
+	graphviz \
+	htop \
+	iputils-ping \
+	libasound2-dev \
+	libatlas-base-dev \
+	libavcodec-dev \
+	libavformat-dev \
+	libboost-dev \
+	libboost-filesystem-dev \
+	libboost-system-dev \
+	libcairo2-dev \
+	libcurl4-openssl-dev \
+	libdc1394-dev \
+	libeigen3-dev \
+	libfaac-dev \
+	libfftw3-dev \
+	libfreetype-dev \
+	libfuse2 \
+	libgdal-dev \
+	libgeos++-dev \
+	libgeos-dev \
+	libgflags-dev \
+	libgirepository1.0-dev \
+	libglib2.0-0 \
+	libgoogle-glog-dev \
+	libgphoto2-dev \
+	libgstreamer-plugins-base1.0-dev \
+	libgstreamer1.0-dev \
+	libgtk-3-dev \
+	libhdf5-dev \
+	libjemalloc-dev \
+	libjpeg-dev \
+	liblapack-dev \
+	libmnl0 \
+	libmp3lame-dev \
+#	libmpdec-dev \
+	libncurses-dev \
+	libnl-3-200 \
+	libnl-3-dev \
+	libnl-route-3-200 \
+	libnl-route-3-dev \
+	libnlopt-dev \
+	libnuma-dev \
+	libopenblas-dev \
+	libopencore-amrnb-dev \
+	libopencore-amrwb-dev \
+	libpci3 \
+	libpng-dev \
+	libportmidi-dev \
+	libproj-dev \
+	libprotobuf-dev \
+	libsdl2-dev \
+	libsdl-image1.2-dev \
+	libsdl-mixer1.2-dev \
+	libsdl-ttf2.0-dev \
+	libsm6 \
+	libsmpeg-dev \
+	libssl-dev \
+	libswscale-dev \
+	libswresample-dev \
+	libtbb-dev \
+	libtheora-dev \
+	libtiff-dev \
+	libv4l-dev \
+	libvorbis-dev \
+	libx264-dev \
+	libxext6 \
+	libxine2-dev \
+	libxml2-dev \
+	libxrender-dev \
+	libxrender1 \
+	libxslt1-dev \
+	libxvidcore-dev \
+	libyaml-dev \
+	libzmq3-dev \
+	lintian \
+	lsof \
+	make \
+	mercurial \
+	media-types \
+	mailcap \
+#	mpich \
+	ncurses-term \
+	net-tools \
+	openjdk-11-jdk \
+	openssh-client \
+	openssh-server \
+	pbzip2 \
+	pciutils \
+	pdsh \
+	pkg-config \
+	proj-bin \
+	protobuf-compiler \
+	pv \
+	python-is-python3 \
+	python3-xkit \
+	python3-gdal \
+	screen-resolution-extra \
+	sudo \
+	subversion \
+	swig \
+	tk \
+	unzip \
+	v4l-utils \
+	vim \
+	x264 \
+	xvfb \
+	xz-utils \
+	sudo \
+	yarn \
+	yasm \
+	zip \
+	tcl \
+	udev
+RUN mkdir -p /opt/oracle && \
+    cd /opt/oracle && \
+    wget https://download.oracle.com/otn_software/linux/instantclient/instantclient-basiclite-linuxx64.zip && \
+    unzip instantclient-basiclite-linuxx64.zip && \
+    rm -f instantclient-basiclite-linuxx64.zip && \
+    cd /opt/oracle/instantclient* && \
+    rm -f *jdbc* *occi* *mysql* *README *jar uidrvci genezi adrci && \
+    echo /opt/oracle/instantclient* > /etc/ld.so.conf.d/oracle-instantclient.conf && \
+    ldconfig && \
+    cd /tmp && \
+    curl -sL https://deb.nodesource.com/setup_20.x -o /tmp/nodesource_setup.sh && \ 
+    bash /tmp/nodesource_setup.sh && \
+    apt update ; apt install nodejs && \
+# Install CUDA + cuDNN
+    mkdir -p /usr/local/nvidia/lib && \
+    ln -s /usr/lib/x86_64-linux-gnu/libcudnn.so.9.6.0 /usr/local/cuda/lib64/libcudnn.so && \
+    ldconfig && \
+    update-alternatives --install /usr/bin/python python /usr/bin/python3 2 
+
+WORKDIR /tmp
+RUN git clone -q --branch=v0.3.27 https://github.com/xianyi/OpenBLAS.git && \
+    cd OpenBLAS && \
+    make DYNAMIC_ARCH=1 NO_AFFINITY=1 NUM_THREADS=48 FC=gfortran && \
+    make install && \
+# install git-lfs
+    cd /tmp && \
+    curl -sLO https://github.com/git-lfs/git-lfs/releases/download/v3.5.1/git-lfs-linux-amd64-v3.5.1.tar.gz && \
+    tar -zxf git-lfs-linux-amd64-v3.5.1.tar.gz && \
+    cd git-lfs-3.5.1 && \
+    bash install.sh && \
+    git clone https://github.com/aristocratos/bashtop.git && \
+    cd bashtop && \
+    make install && \
+    rm -rf /tmp/* 
+
+RUN curl -fL https://github.com/cdr/code-server/releases/download/v$CODESERVER/code-server-$CODESERVER-linux-amd64.tar.gz \
+      | tar -C /usr/local/lib -xz && \ 
+    mv /usr/local/lib/code-server-$CODESERVER-linux-amd64 /usr/local/lib/code-server-$CODESERVER && \
+    ln -s /usr/local/lib/code-server-$CODESERVER/bin/code-server /usr/local/bin/code-server 
+
+# remove hwloc-like packages (ImportError: /opt/hpcx/ucc/lib/libucc.so.1: undefined symbol issue)
+RUN apt-get purge -y hwloc-nox libhwloc-plugins
+# install kss (temporary, when kss package's pyyaml version is updated, move to requirements.txt)
+WORKDIR /tmp
+RUN python3 -m pip install --upgrade pip setuptools wheel Cython
+RUN cd /tmp && git clone https://github.com/hyunwoongko/kss.git && \
+    cd kss && \
+    sed -i 's/pyyaml==6.0/pyyaml>=6.0.2/' setup.py && \
+    pip install . && \
+    rm -rf /tmp/*
+
+COPY ./requirements.24.12.txt /tmp/requirements.txt
+RUN python3 -m pip install --disable-pip-version-check --no-cache-dir -r requirements.txt 
+RUN python3 -m pip install datasets==3.0.1 multiprocess==0.70.16
+
+RUN python3 -m pip install --no-cache-dir \
+      mpi4py>=3.1.4 \
+      nni>=2.10.1 \
+      mlflow>=2.6.0 \
+      scikit-nni>=0.2.1
+
+RUN python3 -m pip install pytorch-lightning && \
+    python3 -m pip install --no-cache-dir \
+        torch-scatter \
+	torch-sparse \
+	torch-cluster \
+	torch-spline-conv \
+	torch-geometric 
+
+
+WORKDIR /tmp
+RUN git clone -q https://github.com/TimDettmers/bitsandbytes.git && \
+    cd /tmp/bitsandbytes && \
+    cmake -B build -DBUILD_CUDA=ON -S . && \
+    python setup.py install
+
+# Install ipython kernelspec
+RUN python3 -m ipykernel install --display-name "PyTorch 2.6 (NGC 24.12/Python 3.12) on Backend.AI" && \
+    cat /usr/local/share/jupyter/kernels/python3/kernel.json
+# Backend.AI specifics
+COPY ./service-defs /etc/backend.ai/service-defs
+COPY ./runner-scripts/bootstrap.sh runner-scripts/setup_multinode.py /opt/container/
+
+LABEL ai.backend.kernelspec="1" \
+      ai.backend.envs.corecount="OPENBLAS_NUM_THREADS,OMP_NUM_THREADS,NPROC" \
+      ai.backend.features="batch query uid-match user-input" \
+      ai.backend.base-distro="ubuntu24.04" \
+      ai.backend.accelerators="cuda" \
+      ai.backend.resource.min.cpu="1" \
+      ai.backend.resource.min.mem="1g" \
+      ai.backend.resource.min.cuda.device=1 \
+      ai.backend.resource.min.cuda.shares=0 \
+      ai.backend.runtime-type="python" \
+      ai.backend.runtime-path="/usr/bin/python" \
+      ai.backend.service-ports="ipython:pty:3000,jupyter:http:8091,jupyterlab:http:8090,vscode:http:8180,tensorboard:http:6006,mlflow-ui:preopen:5000,nniboard:preopen:8080"
+# Install Jupyterlab extensions
+
+RUN python3 -m pip install --no-cache-dir \
+        jupyter_nbextensions_configurator>=0.6.3 \
+	jupyter_contrib_nbextensions>=0.7.0 \
+	jupyter-contrib-core>=0.4.2 \
+	jupyter-highlight-selected-word>=0.2.0 \
+	jupyter-server>=2.14.2 \
+        jupyter_lsp>=2.2.0 
+RUN python3 -m pip install --no-cache-dir \
+	jupyterlab>=4.3.2 \
+	jupyterlab-code-formatter>=3.0.2 \
+	jupyterlab-hdf \
+	jupyterlab-launcher \
+        notebook
+
+RUN python3 -m pip install --no-cache-dir \
+	jupyter-dash \
+	jupyter-js-widgets-nbextension \
+	jupyter-latex-envs \
+	jupyter-packaging
+RUN python3 -m pip install --no-cache-dir \
+	jupyter-server-mathjax \
+	jupyter-server-proxy==1.6.0 \
+	jupyter-telemetry==0.1.0 \
+	jupyter==1.0.0 \
+	jupyter-client==8.4.0 \
+	jupyter_bokeh==2.0.4 
+
+RUN python3 -m pip uninstall -y markupsafe && \
+    python3 -m pip uninstall -y jupyter_core==5.3.0 && \
+    python3 -m pip install markupsafe==2.0.1 &&\
+    python3 -m pip install jsonschema[format,format-nongpl]==4.17.3 
+
+RUN python3 -m pip install widgetsnbextension jupyter_contrib_nbextensions jupyter_lsp markupsafe==2.0.1 && \
+    jupyter lab build --dev-build=False --minimize=False && \
+    jupyter nbextensions_configurator enable && \
+    jupyter contrib nbextension install && \
+    jupyter nbextension enable --py --sys-prefix widgetsnbextension && \
+    jupyter serverextension enable --py jupyterlab --sys-prefix && \
+    jupyter labextension install --no-build @jupyter-widgets/jupyterlab-manager && \
+    jupyter serverextension enable --py jupyter_lsp && \
+    jupyter labextension install --no-build @jupyterlab/toc && \
+    jupyter nbextension enable execute_time/ExecuteTime && \
+    jupyter nbextension enable toc2/main && \
+    jupyter labextension install @jupyterlab/toc && \
+    jupyter lab build    
+
+RUN apt autoclean && \
+    sed -i 's/source \/usr\/local\/nvm\/nvm.sh//' /etc/bash.bashrc && \
+    ln -sf /usr/share/terminfo/x/xterm-color /usr/share/terminfo/x/xterm-256color && \
+    rm -f /tmp/*.whl /tmp/requirem* && \
+    rm -rf /var/lib/apt/lists/* && \	
+    rm -rf /root/.cache && \
+    rm -rf /tmp/*
+
+
+WORKDIR /home/work