faster-rcnn学习之CPU续

1、下载源码

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

2、进入目录lib修改setup.py文件，如下

#CUDA = locate_cuda()

 
 #Extension('nms.gpu_nms',
    #    ['nms/nms_kernel.cu', 'nms/gpu_nms.pyx'],
    #    library_dirs=[CUDA['lib64']],
    #    libraries=['cudart'],
    #    language='c++',
    #    runtime_library_dirs=[CUDA['lib64']],
        # this syntax is specific to this build system
        # we're only going to use certain compiler args with nvcc and not with
        # gcc the implementation of this trick is in customize_compiler() below
    #    extra_compile_args={'gcc': ["-Wno-unused-function"],
    #                        'nvcc': ['-arch=sm_35',
    #                                 '--ptxas-options=-v',
    #                                 '-c',
    #                                 '--compiler-options',
    #                                 "'-fPIC'"]},
    #    include_dirs = [numpy_include, CUDA['include']]
    #),

3、执行make

4、进入目录caffe-fast-rcnn

cp Makefile.config.example Makefile.config

修改Makefile.config 文件，注释掉CUDA相关配置

5、执行

make -j8 && make pycaffe

其中缺少很多头文件，自己安装相应的文件就好

cannot find -lhdf5 参考博客

https://blog.csdn.net/object_allen/article/details/73933934

6、还有一些文件修改参考

https://www.cnblogs.com/justinzhang/p/5386837.html

7、测试

在tools目录下执行python demo.py --cpu

Check failed: *ptr host allocation of size 345600000 failed

貌似是内存不足引起！！！

在一篇博客看到，vgg16是大型模型，需要改成ZF小型模型，或者中型模型

执行./demo.py --cpu --net zf 就可以啦！

8、模型训练

（1）注释tools目录下的tools/train_faster_rcnn_alt_opt.py

 #parser.add_argument('--gpu', dest='gpu_id',
 #                    help='GPU device id to use [0]',
 #                    default=0, type=int)


 # set up caffe
 caffe.set_mode_cpu()#改成cpu模式
 #caffe.set_device(cfg.GPU_ID)

（2）修改experiments/scripts/faster_rcnn_alt_opt.sh

NET=$1
NET_lc=${NET,,}
DATASET=$2

time ../../tools/train_faster_rcnn_alt_opt.py \
  --net_name ${NET} \
  --weights data/imagenet_models/${NET}.v2.caffemodel \
  --imdb ${TRAIN_IMDB} \
  --cfg ../cfgs/faster_rcnn_alt_opt.yml \
  ${EXTRA_ARGS}

set +x
NET_FINAL=`grep "Final model:" ${LOG} | awk '{print $3}'`
set -x

time ../../tools/test_net.py \
  --def models/${PT_DIR}/${NET}/faster_rcnn_alt_opt/faster_rcnn_test.pt \
  --net ${NET_FINAL} \
  --imdb ${TEST_IMDB} \
  --cfg ../cfgs/faster_rcnn_alt_opt.yml \
  ${EXTRA_ARGS}

（3）执行

./faster_rcnn_alt_opt.sh ZF pascal_voc

error1：可以参考博客1；

I0627 10:57:37.710443 10173 solver.cpp:81] Creating training net from train_net file: models/pascal_voc/ZF/faster_rcnn_alt_opt/stage1_rpn_train.pt
F0627 10:57:37.710464 10173 io.cpp:36] Check failed: fd != -1 (-1 vs. -1) File not found: models/pascal_voc/ZF/faster_rcnn_alt_opt/stage1_rpn_train.pt

error2：可以参考博客2；

AttributeError: 'module' object has no attribute 'text_format'

error3：可以参考博客3；

NOT implement！

参考博客1：https://blog.csdn.net/Suii_v5/article/details/73776299

参考博客2：https://blog.csdn.net/qq_33202928/article/details/72526710

参考博客3：https://blog.csdn.net/qq_14975217/article/details/51495844

参考博客4：https://blog.csdn.net/a8039974/article/details/77628805

参考博客5：https://www.cnblogs.com/justinzhang/p/5386837.html

参考博客6：https://blog.csdn.net/object_allen/article/details/73933934

参考博客7：https://blog.csdn.net/Gavin__Zhou/article/details/52052915