MindSpore Flow Installation

View Source On Gitee  

System Environment Information Confirmation

Installation

You can install MindSpore Flow either by pip or by source code.

Installation by pip

export MS_VERSION=2.0.0a0
export MindFlow_VERSION=0.1.0a0
# gpu and ascend are supported
export DEVICE_NAME=gpu
# cuda-10.1 and cuda-11.1 are supported
export CUDA_VERSION=cuda-11.1

# Python3.7
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MS_VERSION}/MindScience/${DEVICE_NAME}/x86_64/${CUDA_VERSION}/mindflow_${DEVICE_NAME}-${MindFlow_VERSION}-cp37-cp37m-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
# Python3.8
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MS_VERSION}/MindScience/${DEVICE_NAME}/x86_64/${CUDA_VERSION}/mindflow_${DEVICE_NAME}-${MindFlow_VERSION}-cp38-cp38-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
# Python3.9
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MS_VERSION}/MindScience/${DEVICE_NAME}/x86_64/${CUDA_VERSION}/mindflow_${DEVICE_NAME}-${MindFlow_VERSION}-cp39-cp39-linux_x86_64.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple

Installation by Source Code

  1. Download source code from Gitee.

    git clone -b r0.3 https://gitee.com/mindspore/mindscience.git
    cd {PATH}/mindscience/MindFlow
    
  2. Compile in Ascend backend.

    bash build.sh -e ascend -j8
    
  3. Compile in GPU backend.

    export CUDA_PATH={your_cuda_path}
    bash build.sh -e GPU -j8
    
  4. Install the compiled .whl file.

    cd {PATH}/mindscience/MindFLow/output
    pip install mindflow_*.whl
    

Installation Verification

Successfully installed, if there is no error message such as No module named 'mindflow' when execute the following command:

python -c 'import mindflow'