win10 安装mmdetection记录

相关环境和依赖：

1.CUDA10.0（即nvcc,下载安装，非conda安装!!）

2.VS2017（下载安装,支持C++编译）

3.pytorch=1.4.0=py3.7_cuda101_cudnn7_0

4.torchvision=0.5.0=py37_cu101

5.cudatoolkit=10.1.243=h74a9793_0

6.mmcv==0.2.16

#################################################################################################

安装pipeline：

1.安装CUDA10.0 （V10.0.130）

https://developer.nvidia.com/cuda-10.0-download-archive?target_os=Windows&target_arch=x86_64&target_version=10&target_type=exelocal

2.安装vs2017（微软官方社区版）

3.安装pytorch的GPU版本（conda安装会自动安装cudatoolkit）

conda create -n mmdet python=3.7 #创建虚拟环境
activate mmdet
conda install pytorch=1.4.0=py3.7_cuda101_cudnn7_0
conda install torchvision=0.5.0=py37_cu101

4.安装cython，pycocotools

pip install cython

Pycocotools 这个东西windows安装需要https://github.com/philferriere/cocoapi下载后进入pythonAPI文件夹内

使用python setup.py install

5.安装mmcv

pip install mmcv==0.2.16

6.安装mmdetection（我这里用的SOLO源码，基于mmdetection1.0.0）

1)下载源码（安装相关依赖）

git clone https://github.com/WXinlong/SOLO
cd SOLO
pip install -r requirements.txt

2）修改编译文件：

1.将~Lib\site-packages\torch\utils\cpp_extension.py 中 ，所有decode() 都改成了decode("utf8","ignore")

2.修改setup.py文件中extra_compile_args 相关代码，增加cxx的："-DMS_WIN64","-MD"
也就是将def make_cuda_ext(name, module, sources) 整个函数改为

def make_cuda_ext(name, module, sources):
 
    return CUDAExtension(
        name='{}.{}'.format(module, name),
        sources=[os.path.join(*module.split('.'), p) for p in sources],
        extra_compile_args={
            'cxx': ["-DMS_WIN64","-MD"],
            'nvcc': [
                '-D__CUDA_NO_HALF_OPERATORS__',
                '-D__CUDA_NO_HALF_CONVERSIONS__',
                '-D__CUDA_NO_HALF2_OPERATORS__',
            ]
        })

3）编译和安装

python setup.py build_ext --inplace
python setup.py install develop

 ###################################################################################################

最后附上环境相关库清单（conda env export > requirements.yaml 导出），仅供参考

name: mmdet
channels:
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch
  - 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/free/
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
  - defaults
dependencies:
  - blas=1.0=mkl
  - ca-certificates=2020.6.24=0
  - certifi=2020.6.20=py37_0
  - cffi=1.14.0=py37h7a1dbc1_0
  - cudatoolkit=10.1.243=h74a9793_0
  - cython=0.29.21=py37ha925a31_0
  - freetype=2.10.2=hd328e21_0
  - icc_rt=2019.0.0=h0cc432a_1
  - intel-openmp=2020.1=216
  - jpeg=9b=hb83a4c4_2
  - libpng=1.6.37=h2a8f88b_0
  - libtiff=4.1.0=h56a325e_1
  - lz4-c=1.9.2=h62dcd97_1
  - mkl=2020.1=216
  - mkl-service=2.3.0=py37hb782905_0
  - mkl_fft=1.1.0=py37h45dec08_0
  - mkl_random=1.1.1=py37h47e9c7a_0
  - msys2-conda-epoch=20160418=1
  - ninja=1.7.2=0
  - numpy=1.18.5=py37h6530119_0
  - numpy-base=1.18.5=py37hc3f5095_0
  - olefile=0.46=py37_0
  - openssl=1.1.1g=he774522_0
  - pip=20.1.1=py37_1
  - pycparser=2.20=py_2
  - python=3.7.7=h81c818b_4
  - pytorch=1.4.0=py3.7_cuda101_cudnn7_0
  - setuptools=49.2.0=py37_0
  - six=1.15.0=py_0
  - sqlite=3.32.3=h2a8f88b_0
  - tk=8.6.10=he774522_0
  - torchvision=0.5.0=py37_cu101
  - vc=14.1=h0510ff6_4
  - vs2015_runtime=14.16.27012=hf0eaf9b_3
  - wheel=0.34.2=py37_0
  - wincertstore=0.2=py37_0
  - xz=5.2.5=h62dcd97_0
  - zlib=1.2.11=h62dcd97_4
  - zstd=1.4.5=h04227a9_0
  - pip:
    - addict==2.2.1
    - albumentations==0.4.6
    - asynctest==0.13.0
    - atomicwrites==1.4.0
    - attrs==19.3.0
    - chardet==3.0.4
    - codecov==2.1.8
    - colorama==0.4.3
    - coverage==5.2.1
    - cycler==0.10.0
    - decorator==4.4.2
    - flake8==3.8.3
    - idna==2.10
    - imagecorruptions==1.1.0
    - imageio==2.9.0
    - imgaug==0.4.0
    - importlib-metadata==1.7.0
    - iniconfig==1.0.0
    - isort==5.2.2
    - jaraco-classes==3.1.0
    - jaraco-collections==3.0.0
    - jaraco-functools==3.0.1
    - jaraco-structures==2.0
    - jaraco-text==3.2.0
    - jaraco-ui==2.0.1
    - jaraco-windows==5.0.0
    - kiwisolver==1.2.0
    - kwarray==0.5.9
    - matplotlib==3.3.0
    - mccabe==0.6.1
    - mmcv==0.2.16
    - mmdet==1.0.0+unknown
    - more-itertools==8.4.0
    - networkx==2.4
    - opencv-python==4.1.2.30
    - ordered-set==4.0.2
    - packaging==20.4
    - pandas==1.1.0
    - path==15.0.0
    - path-py==12.5.0
    - pillow==6.2.2
    - pluggy==0.13.1
    - py==1.9.0
    - pycocotools==2.0
    - pycodestyle==2.6.0
    - pyflakes==2.2.0
    - pyparsing==2.4.7
    - pytest==6.0.1
    - pytest-cov==2.10.0
    - pytest-runner==5.2
    - python-dateutil==2.8.1
    - pytz==2020.1
    - pywavelets==1.1.1
    - pyyaml==5.3.1
    - requests==2.24.0
    - scikit-image==0.17.2
    - scipy==1.5.2
    - shapely==1.7.0
    - terminaltables==3.1.0
    - tifffile==2020.7.24
    - toml==0.10.1
    - ubelt==0.9.1
    - urllib3==1.25.10
    - xdoctest==0.13.0
    - yapf==0.30.0
    - zipp==3.1.0
prefix: D:\Miniconda3\envs\mmdet