diff --git a/spawnerdockerfile/Dockerfile.ngc b/spawnerdockerfile/Dockerfile.ngc index 8d2a8fa..3eb4e6c 100644 --- a/spawnerdockerfile/Dockerfile.ngc +++ b/spawnerdockerfile/Dockerfile.ngc @@ -20,23 +20,16 @@ ENV MLNX_OFED_VERSION=23.10-3.2.2.0 ENV STAGE_DIR=/tmp RUN <&1 -# from https://github.com/huggingface/transformers/blob/main/docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile install deepspeed fail -# reference deepspeed install from https://github.com/microsoft/DeepSpeed/blob/master/docker/Dockerfile # install Mellanox OFED prepare apt-get update -apt install libnvidia-compute-535 +apt install -y libnvidia-compute-535 apt-get install -y pciutils tk kmod libusb-1.0-0 tcl chrpath libpci3 bison lsof graphviz ethtool swig udev libltdl-dev libelf1 libmnl0 debhelper flex libfuse2 # install Mellanox OFED mkdir -p ${STAGE_DIR} -wget -q -O - http://www.mellanox.com/downloads/ofed/MLNX_OFED-${MLNX_OFED_VERSION}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64.tgz | tar xzf - -C ${STAGE_DIR} +wget http://www.mellanox.com/downloads/ofed/MLNX_OFED-${MLNX_OFED_VERSION}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64.tgz -O ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64.tgz +tar xzf ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64.tgz -C ${STAGE_DIR} cd ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64 -./mlnxofedinstall --user-space-only --without-fw-update --all -q +./mlnxofedinstall --user-space-only --without-fw-update --all -q > ${STAGE_DIR}/mlnxofedinstall.log 2>&1 cd ${STAGE_DIR} rm -rf ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64* EOT