search

jwyang / faster-rcnn.pytorch

A faster pytorch implementation of faster r-cnn
MIT License
7.69k stars 2.33k forks source link

why the roi_pooling_forward do nothing?? #391

Open kellenf opened 5 years ago

kellenf commented 5 years ago

I notice you use the roi_pooling_forward and roi_pooling_forward_cuda functions. But in your code it is:

def roi_pooling_forward(pooled_height, pooled_width, spatial_scale, _features, rois, output): return None

def roi_pooling_forward_cuda(pooled_height, pooled_width, spatial_scale, features, rois, output, argmax): return None

......I don't know why it would be this???can you help me? Thank you .(I have not change the code)

haitaobiyao commented 5 years ago

I think you don't choose the roi_pooing,if you not change the code,the code use RCNN_roi_align not roi_pooling.