Teach you how to install PaddleDetection (latest CUDA11.7 version)

foreword

This article records the process of installing PaddleDetection in the linux system, using Conda to install it;

(I tried the docker method, but I couldn’t get the image; I tried the pip method, but I couldn’t find the library; I finally installed it successfully using Conda.)

Table of contents

foreword

1. Set the domestic source to accelerate Conda

2. Create a Conda environment

3. Install PaddlePaddle

4. Install Paddle Detection

---

1. Set the domestic source to accelerate Conda

In the Linux system, configure the source of conda by modifying the condarc file

vim ~/.condarc

 Change it to look like this:

show_channel_urls: true
channel_alias: https://mirrors.tuna.tsinghua.edu.cn/anaconda
default_channels:
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/pro
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/msys2
custom_channels:
  conda-forge: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
  msys2: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
  bioconda: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
  menpo: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
  pytorch: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
  simpleitk: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud

2. Create a Conda environment

Create a Conda environment named PaddleDetection and specify python version 3.8

conda create -n PaddleDetection python=3.8

enter the environment

conda activate PaddleDetection

3. Install PaddlePaddle

First come to the official website, choose the appropriate CUDA version, I chose the latest CUDA11.7

https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/install/conda/linux-conda.html

 Install according to the installation information (command) in the above figure;

conda install paddlepaddle-gpu==2.4.2 cudatoolkit=11.7 -c https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/Paddle/ -c conda-forge

Then configure the environment variable (if not configured, the following error will occur when using it)

dynamic_loader.cc:307] The third-party dynamic library (libcudnn.so) that Paddle depends on is not configured correctly. (error code is /usr/local/cuda/lib64/libcudnn.so: cannot open shared object file: No such file or directory)

Use conda env list to query the path of the environment

(PaddleDetection) root@bap3ac2457:/guopu # conda env list

# conda environments:
#
base                     /opt/conda
PaddleDetection       *  /opt/conda/envs/PaddleDetection

You can see that the path of the newly created Conda is  /opt/conda/envs/PaddleDetection

Add /opt/conda/envs/PaddleDetection/lib/ to environment variables.bashrc

echo "export LD_LIBRARY_PATH=/opt/conda/envs/PaddleDetection/lib/">>~/.bashrc

Test whether the installation is successful

# 确认PaddlePaddle安装成功
python -c "import paddle; paddle.utils.run_check()"

# 确认PaddlePaddle版本
python -c "import paddle; print(paddle.__version__)"

See the successful print message

(PaddleDetection) root@bap3ac2457:/guopu# python -c "import paddle; paddle.utils.run_check()"
Running verify PaddlePaddle program ... 
W0508 08:48:03.937019 44515 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 7.5, Driver API Version: 11.8, Runtime API Version: 11.7
W0508 08:48:03.950474 44515 gpu_resources.cc:91] device: 0, cuDNN Version: 8.4.
PaddlePaddle works well on 1 GPU.
PaddlePaddle works well on 1 GPUs.
PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now.
(PaddleDetection) root@bap3ac2457:/guopu# 

4. Install Paddle Detection

First download the PaddleDetection code

git clone https://github.com/PaddlePaddle/PaddleDetection.git

Enter the project directory

cd PaddleDetection

Use pip to install dependencies, you can use -i https://pypi.tuna.tsinghua.edu.cn/simple to speed up installation

pip install -r requirements.txt

If an error is reported when installing pycocotools, refer to https://blog.csdn.net/weixin_57096837/article/details/122775990

Compile and install paddlelet

python setup.py install

If there is an error downloading some libraries, such as pyclipper, then manually install: pip install pyperclip, and then execute python setup.py install

If error: protobuf 3.20.0 is installed but protobuf>=3.20.2 is required by {'onnx'}, execute: pip install protobuf==3.20.3

test environment

Confirm that the test passes after installation:

python ppdet/modeling/tests/test_architectures.py

After the test is passed, the following information will be displayed:

(PaddleDetection) root@bap3ac2457:/guopu/PaddleDetection# python ppdet/modeling/tests/test_architectures.py
Warning: Unable to use numba in PP-Tracking, please install numba, for example(python3.7): `pip install numba==0.56.4`
Warning: Unable to use numba in PP-Tracking, please install numba, for example(python3.7): `pip install numba==0.56.4`
W0510 00:52:22.579213 47617 gpu_resources.cc:61] Please NOTE: device: 0, GPU Compute Capability: 7.5, Driver API Version: 11.8, Runtime API Version: 11.7
W0510 00:52:22.590267 47617 gpu_resources.cc:91] device: 0, cuDNN Version: 8.4.
.......
----------------------------------------------------------------------
Ran 7 tests in 3.731s

OK
(PaddleDetection) root@bap3ac2457:/guopu/PaddleDetection# 

If the above information is printed, the installation is complete~