时隔半年再战fast-rcnn

Py-Faster-RCNN(Caffe)的安装

1.git clone —recursive https://github.com/rbgirshick/py-faster-rcnn.git

2.sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler libhdf5-dev

sudo apt-get install --no-install-recommends libboost-all-dev
sudo apt-get install libatlas-base-dev
sudo apt-get install python-dev
sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

sudo apt-get install --no-install-recommends libboost-all-dev 

 3. 安装 Cython, easydict, opencv

这三个库一定要记得安装，否则会报错

sudo pip install cython
sudo pip install easydict
sudo apt-get install python-opencv

4.在 /py-faster-rcnn/lib directory 目录下执行make, 会出现如下结果:

5.修改makefile.config

8行

USE_CUDNN := 1

87行

WITH_PYTHON_LAYER := 1

90行

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

其他项，比如atlas、anaconda的python等，看情况修改，不修改也OK。

6。修改makefile

1.LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial

2.LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial

7.

编译caffe-fast-rcnn

需要先更新用到的caffe：原版py-faster-rcnn依赖的caffe比较老，不支持cudnnv5,那就更新下对应的源码。提供两种版本，方法1更简单些。

利用git修改 参考https://www.cnblogs.com/zf-blog/p/6701710.html

http://blog.csdn.net/10km/article/details/62418583

cd caffe-fast-rcnn  
git remote add caffe https://github.com/BVLC/caffe.git  
git fetch caffe  
git merge -X theirs caffe/master  

然后在合并之后注释掉include/caffe/layers/python_layer.hpp文件里的self_.attr(“phase”) = static_cast(this->phase_)

 参考：http://blog.csdn.net/nicky_lyu/article/details/53181434

https://www.cnblogs.com/zjutzz/p/6034408.html

make runtest的时候出现 Check failed: error == cudaSuccess (30 vs. 0) unknown error

（网上也有说把makefile.config文件中的compute改了 我的显卡是gtx 1050 照着官网改成了61改完再加sudo就成功了 估计不用改）

sudo make runtest ok！