Open ccyxx opened 7 years ago
@ccyxx Has the problem been solved? My question is the same as yours Thank you
@punkyoon hello your head_trainval.prototxt is Python type,and you .sh uses c++ and get eror "No module named roi_data_layer". how to solve this problem thank you
The author used his own caffe and he didn't mention it in this readme https://github.com/yokattame/caffe
@Smellly thanks
@Smellly thank you
@yokattame hello I just train your base network in github and get the better result than you report in paper considerably. for example market1501 expecially, I get the 79.4% in top-1 and paper reports is 76.9%(Spindle Net) and 72.1%(full part) respectively. In addation, The base network result in Prid and Senserid datasets also better than Spindel Net's respectively. what's the reason for this phenomenon? Thanks very much! I am in the Hello World phase in re-id.
@liuajian can you share the trained base network model with me? Thanks.
@suiyuejinghao1234567 because the image path is wrong in roi_data_layer.py, you need to change the code ' image_path = self.root_folder + f.readline().strip() ' to ' image_path = f.readline().strip() '
I1102 11:24:54.396538 23029 layer_factory.hpp:77] Creating layer data/head ImportError: No module named roi_data_layer what shoud I do? @liuajian @Smellly
@Smellly How to install caffe in https://github.com/yokattame/caffe?
when I run "./scripts/train_head.sh", get error "No module named roi_data_layer".
ps:my caffe version:"https://github.com/rbgirshick/py-faster-rcnn"