Ssd.pytorch: AttributeError: 'NoneType' object has no attribute 'shape'
I DON'T KNOW WHY AND HOW TO FIX IT.
LAWSSSS
All 4 comments
I have the same error.
Brozhao
on 28 Sep 2018
I have the same error, too. I have spent two days solving it. Maybe the format of your dataset is wrong. My dataset has 9 images whose format is .JPG while others are .jpg, so I correct them to .jpg, and then I overcome this error.
pxjx
on 12 Oct 2018
I have the same error, too. I have spent two days solving it. Maybe the format of your dataset is wrong. My dataset has 9 images whose format is .JPG while others are .jpg, so I correct them to .jpg, and then I overcome this error.
Thanks. I'll give it a try. Best regards.
LAWSSSS
on 22 Oct 2018
You have to make sure that your images are in the directory structure as VOC.
Check voc0712.py in VOCDetection
self._imgpath = osp.join('%s', 'JPEGImages', '%s.jpg')
You can also change this as per need.
adityac8
on 12 Mar 2019
Related issues
Elyorcv
路
4Comments
lmingyin
路
5Comments
FostorHUNT
路
4Comments
mikejmills
路
6Comments
oarriaga
路
4Comments
Most helpful comment
You have to make sure that your images are in the directory structure as VOC.
Check
voc0712.pyinVOCDetectionself._imgpath = osp.join('%s', 'JPEGImages', '%s.jpg')You can also change this as per need.