安装深度学习环境

玄云飘风

已于 2025-03-05 08:48:05 修改

阅读量229

点赞数

CC 4.0 BY-SA版权

分类专栏：装机/系统/环境文章标签：计算机视觉深度学习

于 2018-03-26 11:05:34 首次发布

本文链接：https://blog.youkuaiyun.com/tfcy694/article/details/79695052

装机/系统/环境专栏收录该内容

14 篇文章

订阅专栏

本文提供在Ubuntu 20.04上安装CUDA 11.7及11.8版本的详细步骤，并指导如何配置Python环境以支持PyTorch 2.0.1。此外，还介绍了安装MMDetection和MMTracking所需的依赖项及验证安装成功的方法。

OS: ubuntu20.04

CUDA

11.7.1
https://developer.nvidia.com/cuda-11-7-1-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=deb_local
https://docs.nvidia.com/cuda/archive/11.7.1/cuda-installation-guide-linux/
https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin
https://developer.download.nvidia.com/compute/cuda/11.7.1/local_installers/cuda-repo-ubuntu2004-11-7-local_11.7.1-515.65.01-1_amd64.deb

11.8.0
https://developer.nvidia.com/cuda-11-8-0-download-archive?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=20.04&target_type=deb_local
https://docs.nvidia.com/cuda/archive/11.8.0/cuda-installation-guide-linux/
https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin
https://developer.download.nvidia.com/compute/cuda/11.8.0/local_installers/cuda-repo-ubuntu2004-11-8-local_11.8.0-520.61.05-1_amd64.deb

# 验证
git clone git@github.com:NVIDIA/cuda-samples.git -b v11.8

cuDNN

8.9.0
https://developer.nvidia.com/rdp/cudnn-archive
https://developer.nvidia.com/downloads/compute/cudnn/secure/8.9.0/local_installers/11.8/cudnn-local-repo-ubuntu2004-8.9.0.131_1.0-1_amd64.deb
https://docs.nvidia.com/deeplearning/cudnn/archives/cudnn-890/install-guide/

python

# https://repo.anaconda.com/miniconda/Miniconda3-py310_24.1.2-0-Linux-x86_64.sh
sh Miniconda3-py310_24.1.2-0-Linux-x86_64.sh
pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117  --extra-index-url https://download.pytorch.org/whl/cu117
pip install torch==2.0.1 torchvision==0.15.2

python3 -c "import torch;print(torch.cuda.is_available())"

mmdet

# CUDA: 11.8, python: 3.10, PyTorch: 2.0.1
pip install cuda-python

# 安装mmcv
pip3 install mmengine
pip3 install mmcv -f https://download.openmmlab.com/mmcv/dist/cu118/torch2.0.0/index.html
# 在 MMCV1.x 中，它有两个包名 mmcv-full 和 mmcv，前者比后者多包含了 CUDA 算子。在 MMCV2.x 中，我们决定将 mmcv-full 重命名为 mmcv，将 mmcv 重名为 mmcv-lite。
# https://zhuanlan.zhihu.com/p/572550592

# 安装mmdet
git clone https://github.com/open-mmlab/mmdetection.git -b v3.3.0
cd mmdetection
pip3 install -e .

# 安装mmtracking
git clone https://github.com/open-mmlab/mmtracking.git -b v1.0.0rc1
cd mmtracking
pip3 install -r requirements/build.txt
pip3 install -v -e .  # or "python3 setup.py develop"
# 可选数据集依赖
pip3 install git+https://github.com/JonathonLuiten/TrackEval.git
pip3 install git+https://github.com/lvis-dataset/lvis-api.git
pip3 install git+https://github.com/TAO-Dataset/tao.git
pip3 install git+https://github.com/votchallenge/toolkit.git

# CUDA: 11.7, python: 3.8, PyTorch: 2.0.1
pip install cuda-python

# 安装mmdet
# 方法一
pip3 install mmengine
pip3 install mmcv -f https://download.openmmlab.com/mmcv/dist/cu117/torch2.0.0/index.html
# 在 MMCV1.x 中，它有两个包名 mmcv-full 和 mmcv，前者比后者多包含了 CUDA 算子。在 MMCV2.x 中，我们决定将 mmcv-full 重命名为 mmcv，将 mmcv 重名为 mmcv-lite。
# https://zhuanlan.zhihu.com/p/572550592
git clone https://github.com/open-mmlab/mmdetection.git -b v3.3.0
cd mmdetection
pip3 install -e .
# 方法二
pip install -U openmim
mim install mmengine
mim install "mmcv==2.0.0"
git clone https://github.com/open-mmlab/mmdetection.git -b v3.3.0
cd mmdetection
pip install -v -e .

# 安装mmsegmentation
git clone -b v1.2.2 https://github.com/open-mmlab/mmsegmentation.git
cd mmsegmentation
pip install -v -e .

# 安装mmtracking
git clone https://github.com/open-mmlab/mmtracking.git -b v1.0.0rc1
cd mmtracking
pip3 install -r requirements/build.txt
pip3 install -v -e .  # or "python3 setup.py develop"
# 可选数据集依赖
pip3 install git+https://github.com/JonathonLuiten/TrackEval.git
pip3 install git+https://github.com/lvis-dataset/lvis-api.git
pip3 install git+https://github.com/TAO-Dataset/tao.git
pip3 install git+https://github.com/votchallenge/toolkit.git

验证安装

mim download mmsegmentation --config pspnet_r50-d8_4xb2-40k_cityscapes-512x1024 --dest .
python demo/image_demo.py demo/demo.png configs/pspnet/pspnet_r50-d8_4xb2-40k_cityscapes-512x1024.py pspnet_r50-d8_512x1024_40k_cityscapes_20200605_003338-2966598c.pth --device cuda:0 --out-file result.jpg

mim download mmdet --config rtmdet_tiny_8xb32-300e_coco --dest .
python demo/image_demo.py demo/demo.jpg rtmdet_tiny_8xb32-300e_coco.py --weights rtmdet_tiny_8xb32-300e_coco_20220902_112414-78e30dcc.pth --device cpu

python3 -c "import mmdet;print(mmdet.__version__)"
python3 demo/demo_mot_vis.py configs/mot/deepsort/sort_faster-rcnn_fpn_4e_mot17-private.py --input demo/demo.mp4 --output mot.mp4

ref:
https://github.com/open-mmlab/mmdetection
https://mmdetection.readthedocs.io/en/latest/get_started.html
https://github.com/open-mmlab/mmtracking/blob/master/docs/en/install.md