Skip to content

Latest commit

 

History

History
47 lines (39 loc) · 1.54 KB

INSTALL.md

File metadata and controls

47 lines (39 loc) · 1.54 KB

Install CUDA

conda install --channel "nvidia/label/cuda-12.1.0" cuda-toolkit

Install PyTorch:

conda install pytorch torchvision torchaudio pytorch-cuda==12.1 -c pytorch -c nvidia

Or install with CUDA11.8

conda install pytorch torchvision torchaudio cuda-toolkit pytorch-cuda==11.8 -c pytorch -c "nvidia/label/cuda-11.8.0"

Install mmengine:

python -m pip install https://github.com/open-mmlab/mmengine/archive/refs/tags/v0.8.5.zip

Install mmcv:

TORCH_CUDA_ARCH_LIST="8.0" TORCH_NVCC_FLAGS="-Xfatbin -compress-all" CUDA_HOME=$(dirname $(dirname $(which nvcc))) LD_LIBRARY_PATH=$(dirname $(dirname $(which nvcc)))/lib MMCV_WITH_OPS=1 FORCE_CUDA=1 python -m pip install git+https://github.com/open-mmlab/mmcv.git@4f65f91db6502d990ce2ee5de0337441fb69dd10

also try with

MMCV_CUDA_ARGS="-std=c++17" 

when using mmcv 1.x with Pytorch 1.12.1 or later version

Install other openmmlab packages:

python -m pip install \
https://github.com/open-mmlab/mmdetection/archive/refs/tags/v3.1.0.zip \
https://github.com/open-mmlab/mmsegmentation/archive/refs/tags/v1.1.1.zip \
https://github.com/open-mmlab/mmpretrain/archive/refs/tags/v1.0.1.zip

Install extra packages:

python -m pip install git+https://github.com/cocodataset/panopticapi.git \
git+https://github.com/HarborYuan/lvis-api.git \
tqdm terminaltables pycocotools scipy tqdm ftfy regex timm scikit-image kornia

THe default environment is openmmlab on Shanghai AI server.