Tian Zhi

@tengshaofeng Please try to compile it with CUDA 7.0. Thank you:-).

@tengshaofeng You can refer to https://github.com/BVLC/caffe/issues/410.

@yingning I don't know why you have this error. But please try to run `export PYTHONPATH=src/:$PYTHONPATH` before `python tools/demo.py` in the same working directory. Thank you.

Please try to run `export PYTHONPATH=src/:$PYTHONPATH` before running the command line. Thank you.

@rajma996 I google for this error and find it may have to do with the version of CUDNN. Please double check that the caffe in this program was compiled with...

@zq2323 the ImageNet pre-trained VGG can be downloaded here http://www.robots.ox.ac.uk/~vgg/software/very_deep/caffe/VGG_ILSVRC_16_layers.caffemodel.

@chaobeiying You could try this fork (https://github.com/qingswu/CTPN). Thank you very much!

@dhimanpd thank you for your interest. The training scripts are not released due to the cooperation with industry.

@anguoyang If the problem is because of cpu_nms.pyx, I suggest you could use other nms implementation in python instead. You can find a branch of them on github. Thank you.

@guangdongliang What are the margins?