Install on Linux via PIP¶

1.1 PREQUISITES¶

• Linux Version (64 bit)

  • CentOS 7 (GPUVersion Supports CUDA 10.1/10.2/11.0/11.1/11.2)**

  • Ubuntu 16.04 (GPUVersion Supports CUDA 10.1/10.2/11.0/11.1/11.2)

  • Ubuntu 18.04 (GPUVersion Supports CUDA 10.1/10.2/11.0/11.1/11.2)

  • Ubuntu 20.04 (GPUVersion Supports CUDA 10.1/10.2/11.0/11.1/11.2)

• Python Version: 3.6/3.7/3.8/3.9 (64 bit)

• pip or pip3 Version 20.2.2 or above (64 bit)

1.2 How to check your environment¶

• You can use the following commands to view the local operating system and bit information

  uname -m && cat /etc/*release
  

• Confirm that the Python where you need to install PaddlePaddle is your expected location, because your computer may have multiple Python

  • Use the following command to output Python path. Depending on the environment, you may need to replace Python in all command lines in the description with specific Python path

    which python
    

• You need to confirm whether the version of Python meets the requirements

  • Use the following command to confirm that it is 3.6/3.7/3.8/3.9

    python --version
    

• It is required to confirm whether the version of pip meets the requirements. The version of pip is required to be 20.2.2 or above

  python -m ensurepip
  

  python -m pip --version
  

• You need to confirm that Python and pip are 64bit, and the processor architecture is x86_64(or called x64、Intel 64、AMD64). Currently, paddlepaddle does not support arm64 architecture. The first line below outputs “64bit”, and the second line outputs “x86_64”, “x64” or “AMD64”

  python -c "import platform;print(platform.architecture()[0]);print(platform.machine())"
  

• The installation package provided by default requires computer support for MKL

• If you do not know the machine environment, please download and useQuick install script, for instructions please refer tohere。

Choose CPU/GPU¶

• If your computer doesn’t have NVIDIA® GPU, please install the CPU Version of PaddlePaddle

• If your computer has NVIDIA® GPU, please make sure that the following conditions are met and install the GPU Version of PaddlePaddle

  • CUDA toolkit 10.1/10.2 with cuDNN 7 (cuDNN version>=7.6.5, for multi card support, NCCL2.7 or higher)

  • CUDA toolkit 11.0 with cuDNN v8.0.4(for multi card support, NCCL2.7 or higher)

  • CUDA toolkit 11.1 with cuDNN v8.1.1(for multi card support, NCCL2.7 or higher)

  • CUDA toolkit 11.2 with cuDNN v8.1.1(for multi card support, NCCL2.7 or higher)

  • Hardware devices with GPU computing power over 3.5

    You can refer to NVIDIA official documents for installation process and configuration method of CUDA and cudnn. Please refer to CUDA，cuDNN

• If you need to use a multi-card environment, please make sure that you have installed nccl2 correctly, or install nccl2 according to the following instructions (here are the installation instructions of nccl2 under CUDA10.2 and cuDNN7. For more version installation information, please refer to NVIDIA Official Website):

  • Centos system can refer to the following commands

    wget http://developer.download.nvidia.com/compute/machine-learning/repos/rhel7/x86_64/nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm
    

    rpm -i nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm
    

    yum update -y
    

    yum install -y libnccl-2.7.8-1+cuda10.2 libnccl-devel-2.7.8-1+cuda10.2 libnccl-static-2.7.8-1+cuda10.2
    

  • Ubuntu system can refer to the following commands

    wget https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64/nvidia-machine-learning-repo-ubuntu1604_1.0.0-1_amd64.deb
    

    dpkg -i nvidia-machine-learning-repo-ubuntu1604_1.0.0-1_amd64.deb
    

    sudo apt install -y libnccl2=2.7.8-1+cuda10.2 libnccl-dev=2.7.8-1+cuda10.2
    

2.1 CPU Version of PaddlePaddle¶

python -m pip install paddlepaddle==2.2.2 -i https://mirror.baidu.com/pypi/simple

2.2 GPU Version of PaddlePaddle¶

2.2.2 If you are using CUDA 10.1

python -m pip install paddlepaddle-gpu==2.2.2.post101 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html

2.2.2 If you are using CUDA 10.2

python -m pip install paddlepaddle-gpu==2.2.2 -i https://mirror.baidu.com/pypi/simple

2.2.3 If you are using CUDA 11.0

python -m pip install paddlepaddle-gpu==2.2.2.post110 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html

2.2.4 If you are using CUDA 11.1

python -m pip install paddlepaddle-gpu==2.2.2.post111 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html

2.2.5 If you are using CUDA 11.2

python -m pip install paddlepaddle-gpu==2.2.2.post112 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/avx/stable.html

Note：

• If you are using ampere-based GPU, CUDA 11 above version is recommended; otherwise CUDA 10.2 is recommended for better performance.

• Please confirm that the Python where you need to install PaddlePaddle is your expected location, because your computer may have multiple Python. Depending on the environment, you may need to replace Python in all command lines in the instructions with Python 3 or specific Python path.

• If you want to use the tsinghua pypi, you can use the following command:

   python -m pip install paddlepaddle-gpu==[Version] -i https://pypi.tuna.tsinghua.edu.cn/simple
  

• The above commands install the avx package by default. If your machine does not support avx, you need to install the Paddle package of noavx, you can use the following command to install，noavx version paddle wheel only support python3.8：

  First use the following command to download the wheel package to the local, and then use python -m pip install [name].whl to install locally ([name] is the name of the wheel package):

  • cpu and mkl version installed on noavx machine：

  python -m pip download paddlepaddle==2.2.2 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/noavx/stable.html --no-index --no-deps
  

  • cpu and openblas version installed on noavx machine：

  python -m pip download paddlepaddle==2.2.2 -f https://www.paddlepaddle.org.cn/whl/linux/openblas/noavx/stable.html --no-index --no-deps
  

  • GPU cuda10.1 version install on noavx machine：

  python -m pip download paddlepaddle-gpu==2.2.2.post101 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/noavx/stable.html --no-index --no-deps
  

  • GPU cuda10.2 version install on noavx machine：

  python -m pip download paddlepaddle-gpu==2.2.2 -f https://www.paddlepaddle.org.cn/whl/linux/mkl/noavx/stable.html --no-index --no-deps
  

  To determine whether your machine supports avx, you can use the following command. If the output contains avx, it means that the machine supports avx:

  cat /proc/cpuinfo | grep -i avx
  

• If you want to install the Paddle package with avx and openblas, you can use the following command to download the wheel package to the local, and then use python -m pip install [name].whl to install locally ([name] is the name of the wheel package):

  python -m pip download paddlepaddle==2.2.2 -f https://www.paddlepaddle.org.cn/whl/linux/openblas/avx/stable.html --no-index --no-deps