Creating object detection dataset throw an attribute error

Hello,

I am using NGC DIGITS nvcr.io/nvidia/digits:20.02-tensorflow-py3

I wanted to try object detection by first create an object detection dataset.

I have a dataset of images and txt files (for labels).

Label example: apple_77.txt

apple 0.0 0 0 71 60 175 164 0 0 0 0 0 0 0 0
apple 0.0 0 0 12 22 105 111 0 0 0 0 0 0 0 0
apple 0.0 0 0 134 23 243 115 0 0 0 0 0 0 0 0
apple 0.3 1 0 107 126 216 229 0 0 0 0 0 0 0 0
apple 0.3 1 0 207 138 298 229 0 0 0 0 0 0 0 0

But the problem is that when I press the “create” button after configuring my dataset an error page appears:

AttributeError

‘_csv.reader’ object has no attribute ‘next’

Any idea why? Is it possible to fix it?

Thanks.

Could you try this ?
https://github.com/NVIDIA/DIGITS/issues/2212