Install on Windows via PIP¶

1.1 How to check your environment¶

• Confirm whether the Python version meets the requirements

  • Use the following command to confirm that it is 3.6+/3.7+/3.8+/3.9+/3.10+

    python --version
    

• 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). 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

• NCCL, distribution are not supported on windows now

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 10.2 with cuDNN v7.6.5(for PaddleTensorRT deployment, TensorRT7.0.0.11)

  • CUDA toolkit 11.2 with cuDNN v8.2.1(for PaddleTensorRT deployment, TensorRT8.2.4.2)

  • CUDA toolkit 11.6 with cuDNN v8.4.0(for PaddleTensorRT deployment, TensorRT8.4.0.6)

  • CUDA toolkit 11.7 with cuDNN v8.4.1(for PaddleTensorRT deployment, TensorRT8.4.2.4)

  • GPU CUDA capability over 3.5

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

2.1 CPU Version of PaddlePaddle¶

python -m pip install paddlepaddle==2.4.1 -i https://pypi.tuna.tsinghua.edu.cn/simple

2.2 GPU Version of PaddlePaddle¶

2.2.1 If you are using CUDA 10.2

python -m pip install paddlepaddle-gpu==2.4.1 -i https://pypi.tuna.tsinghua.edu.cn/simple

2.2.2 If you are using CUDA 11.2

python -m pip install paddlepaddle-gpu==2.4.1.post112 -f https://www.paddlepaddle.org.cn/whl/windows/mkl/avx/stable.html

2.2.3 If you are using CUDA 11.6

python -m pip install paddlepaddle-gpu==2.4.1.post116 -f https://www.paddlepaddle.org.cn/whl/windows/mkl/avx/stable.html

2.2.4 If you are using CUDA 11.7

python -m pip install paddlepaddle-gpu==2.4.1.post117 -f https://www.paddlepaddle.org.cn/whl/windows/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 specific Python path.

• 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==2.4.1 -f https://www.paddlepaddle.org.cn/whl/windows/mkl/noavx/stable.html --no-index --no-deps
  

  • cpu and openblas version installed on noavx machine：

  python -m pip download paddlepaddle==2.4.1 -f https://www.paddlepaddle.org.cn/whl/windows/openblas/noavx/stable.html --no-index --no-deps
  

  • GPU cuda10.2 version install on noavx machine：

  python -m pip download paddlepaddle-gpu==2.4.1 -f https://www.paddlepaddle.org.cn/whl/windows/mkl/noavx/stable.html --no-index --no-deps
  

  To determine whether your machine supports avx, you can install the CPU-Z tool to view the “processor-instruction set”.

• 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==2.4.1 -f https://www.paddlepaddle.org.cn/whl/windows/openblas/avx/stable.html --no-index --no-deps