search

Azure / ObjectDetectionUsingCntk

Fast R-CNN Object Detection on Azure using CNTK
129 stars 61 forks source link

No module named cython_bbox #4

Closed AkshayAgarwal007 closed 7 years ago

AkshayAgarwal007 commented 7 years ago

I am trying this out on Azure Linux Data Science VM. I am getting the following error when I am executing the A1_annotateImages.py file.

Traceback (most recent call last): File "A1_annotateImages.py", line 2, in import PARAMETERS File "/home/AkshayAgarwal007/Desktop/detection/PARAMETERS.py", line 1, in from cntk_helpers import * File "/home/AkshayAgarwal007/Desktop/detection/cntk_helpers.py", line 5, in from fastRCNN.nms import nms as nmsPython File "/home/AkshayAgarwal007/Desktop/detection/fastRCNN/init.py", line 7, in from .imdb import imdb File "/home/AkshayAgarwal007/Desktop/detection/fastRCNN/imdb.py", line 15, in from utils_compiled64bit.cython_bbox import bbox_overlaps ImportError: No module named cython_bbox

Seems to be an import issue with .pyd file. How to resolve this?

PatrickBue commented 7 years ago

You are getting this error since the 'cython_bbox' binaries are compiled for windows. A quick fix would be to switch to a VM which runs windows. Alternately, you could install and compile for Linux yourself. See the original fast R-CNN code for that, e.g. https://github.com/rbgirshick/fast-rcnn/tree/master/lib/utils

AkshayAgarwal007 commented 7 years ago

Thanks. I am using the Azure Windows Data Science VM now. It's working fine.