update to ngc images
This commit is contained in:
Your Name
134
spawnerdockerfile/Dockerfile.ngc
Normal file
134
spawnerdockerfile/Dockerfile.ngc
Normal file
@@ -0,0 +1,134 @@
|
||||
ARG REGISTRY=quay.io
|
||||
ARG OWNER=jupyter
|
||||
ARG LABEL=notebook
|
||||
ARG VERSION
|
||||
ARG BASE_CONTAINER=$REGISTRY/$OWNER/$LABEL:$VERSION
|
||||
FROM $BASE_CONTAINER
|
||||
ARG HTTP_PROXY
|
||||
ARG HTTPS_PROXY
|
||||
ENV http_proxy=${HTTP_PROXY}
|
||||
ENV https_proxy=${HTTPS_PROXY}
|
||||
ARG DEBIAN_FRONTEND="noninteractive"
|
||||
ENV DEBIAN_FRONTEND=${DEBIAN_FRONTEND}
|
||||
ARG ROOT_PASSWD="root"
|
||||
ENV ROOT_PASSWD=${ROOT_PASSWD}
|
||||
WORKDIR /root
|
||||
SHELL ["/bin/bash", "-c"]
|
||||
|
||||
# base tools
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
apt-get update
|
||||
apt-get install -y bash-completion wget curl htop jq vim bash libaio-dev build-essential openssh-server python3 python3-pip bzip2 sudo
|
||||
apt-get install -y --no-install-recommends software-properties-common build-essential autotools-dev nfs-common pdsh cmake g++ gcc curl wget vim tmux emacs less unzip htop iftop iotop ca-certificates openssh-client openssh-server rsync iputils-ping net-tools sudo llvm-dev re2c
|
||||
add-apt-repository ppa:git-core/ppa -y
|
||||
apt-get install -y git libnuma-dev wget
|
||||
# Configure SSH for password and public key authentication
|
||||
sed -i 's/#PermitRootLogin prohibit-password/PermitRootLogin yes/' /etc/ssh/sshd_config
|
||||
sed -i 's/#PasswordAuthentication yes/PasswordAuthentication yes/' /etc/ssh/sshd_config
|
||||
sed -i 's/PubkeyAuthentication no/PubkeyAuthentication yes/' /etc/ssh/sshd_config
|
||||
sed -i 's/^#Port 22/Port 22/' /etc/ssh/sshd_config
|
||||
sed -i 's/^Port [0-9]*/Port 22/' /etc/ssh/sshd_config
|
||||
mkdir /var/run/sshd
|
||||
echo "root:${ROOT_PASSWD}" | chpasswd
|
||||
mkdir -p ~/.pip
|
||||
# install miniconda
|
||||
wget -qO- https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O /tmp/miniconda.sh
|
||||
bash /tmp/miniconda.sh -b -p /opt/conda
|
||||
rm /tmp/miniconda.sh
|
||||
ln -s /opt/conda/etc/profile.d/conda.sh /etc/profile.d/conda.sh
|
||||
echo ". /opt/conda/etc/profile.d/conda.sh" >> ~/.bashrc
|
||||
. /opt/conda/etc/profile.d/conda.sh
|
||||
conda init bash
|
||||
conda config --set show_channel_urls true
|
||||
# 配置 .condarc 文件
|
||||
cat <<EOF > ~/.condarc
|
||||
channels:
|
||||
- conda-forge
|
||||
- bioconda
|
||||
- pytorch
|
||||
- pytorch-nightly
|
||||
- nvidia
|
||||
- defaults
|
||||
show_channel_urls: true
|
||||
EOF
|
||||
# install pixi
|
||||
curl -fsSL https://pixi.sh/install.sh | bash
|
||||
EOT
|
||||
|
||||
ENV STAGE_DIR=/tmp
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
pip install -v -U git+https://github.com/facebookresearch/xformers.git@main#egg=xformers
|
||||
pip install git+https://github.com/huggingface/transformers
|
||||
EOT
|
||||
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
git clone https://github.com/microsoft/DeepSpeed-Kernels.git ${STAGE_DIR}/DeepSpeed-Kernels
|
||||
cd ${STAGE_DIR}/DeepSpeed-Kernels
|
||||
python -m pip install -v .
|
||||
EOT
|
||||
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
git clone https://github.com/oneapi-src/oneCCL.git ${STAGE_DIR}/oneCCL
|
||||
cd ${STAGE_DIR}/oneCCL
|
||||
git checkout .
|
||||
git checkout master
|
||||
mkdir build
|
||||
cd build
|
||||
cmake .. -DCMAKE_INSTALL_PREFIX=/usr/local
|
||||
make -j"$(nproc)" install
|
||||
EOT
|
||||
|
||||
ARG DEEPSPEED_VERSION="v0.14.3"
|
||||
ENV DEEPSPEED_VERSION=${DEEPSPEED_VERSION}
|
||||
ARG DEEPSPEED_INSTALL_FLAGS="--allow_sudo --pip_sudo --verbose"
|
||||
ENV DEEPSPEED_INSTALL_FLAGS=${DEEPSPEED_INSTALL_FLAGS}
|
||||
ARG DS_BUILD_SPARSE_ATTN=0
|
||||
ENV DS_BUILD_SPARSE_ATTN=${DS_BUILD_SPARSE_ATTN}
|
||||
ARG DS_BUILD_FUSED_ADAM=1
|
||||
ENV DS_BUILD_FUSED_ADAM=${DS_BUILD_FUSED_ADAM}
|
||||
ARG DS_BUILD_CPU_ADAM=1
|
||||
ENV DS_BUILD_CPU_ADAM=${DS_BUILD_CPU_ADAM}
|
||||
ARG DS_BUILD_OPS=1
|
||||
ENV DS_BUILD_OPS=${DS_BUILD_OPS}
|
||||
ARG HOSTFILE_CONTENT=""
|
||||
ENV HOSTFILE_CONTENT=${HOSTFILE_CONTENT}
|
||||
ENV CUTLASS_PATH="/opt/pytorch/pytorch/third_party/cutlass"
|
||||
ENV CUDA_HOME="/usr/local/cuda"
|
||||
ENV LD_LIBRARY_PATH=${CUDA_HOME}/lib64:${LD_LIBRARY_PATH}
|
||||
ENV PATH=${CUDA_HOME}/bin:${PATH}
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
git clone https://github.com/microsoft/DeepSpeed.git ${STAGE_DIR}/DeepSpeed
|
||||
cd ${STAGE_DIR}/DeepSpeed
|
||||
git checkout ${DEEPSPEED_VERSION}
|
||||
./install.sh ${DEEPSPEED_INSTALL_FLAGS}
|
||||
ds_report
|
||||
EOT
|
||||
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
python -m pip install --upgrade pip
|
||||
python -m pip install peft tiktoken seaborn diffusers blobfile open_clip_torch zstandard mpi4py
|
||||
# python -m pip install --no-deps git+https://github.com/huggingface/optimum.git
|
||||
EOT
|
||||
|
||||
RUN <<EOT
|
||||
#!/bin/bash
|
||||
# 项目目录中的定义通常会覆盖用户家目录中的定义
|
||||
# 配置 .deepspeed_env 文件
|
||||
cat <<EOF > ~/.deepspeed_env
|
||||
TORCH_USE_CUDA_DSA=1
|
||||
DEEPSPEED_VERBOSE=1
|
||||
DEEPSPEED_LOG_LEVEL=DEBUG
|
||||
CUTLASS_PATH=${CUTLASS_PATH}
|
||||
TORCH_CUDA_ARCH_LIST=${TORCH_CUDA_ARCH_LIST}
|
||||
CUDA_HOME=${CUDA_HOME}
|
||||
LD_LIBRARY_PATH=${LD_LIBRARY_PATH}
|
||||
EOF
|
||||
EOT
|
||||
|
||||
CMD ["/usr/sbin/sshd", "-D"]
|
||||
Reference in New Issue
Block a user