py-faster-rcnn安装及训练自己的数据集

一、硬件和环境：

1. GTX 1080 Ti, Cuda 9.0, Cudnn 7.6.0
2. Ubuntu16.04
3. Anaconda3 caffe3虚拟环境
   （caffe3虚拟环境用pip install opencv-python和pip install opencv-contrib-python，然后下载caffe并成功编译。）
4. https://blog.csdn.net/la_fe_/article/details/88547816
5. https://github.com/rbgirshick/py-faster-rcnn

二、下载 py-faster-rcnn并修改Makefile.config文件

source activate caffe3
cd  workspace/faster-rcnn/
git clone --recursive https://github.com/rbgirshick/py-faster-rcnn.git
cd py-faster-rcnn/
cd lib/
make
cd ..
cd caffe-fast-rcnn/
vi Makefile.config

将在workspace/caffe/Makefile.config和Makefile文件复制到workspace/faster-rcnn/py-faster-rcnn/caffe-fast-rcnn/文件夹下，其中Makefile不改变，而Makefile.config文件根据视频里面的Makefile.config文件做一些修改，如

1. 注释掉USE_CUDNN：=1
2. 将

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib 

改为：

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu/hdf5/serial

三、修改Makefile

vi Makefile

四、编译py-faster-rcnn

make -j8
make pycaffe

打开workspace/faster-rcnn/py-faster-rcnn/data/scripts/fetch_faster_rcnn_models.sh，复制里面的网址https://dl.dropboxusercontent.com/s/o6ii098bu51d139/faster_rcnn_models.tgz?dl=0下载faster-rcnn_models
下载好之后把模型放在py-faster-rcnn/data/文件夹下。