Undefined behaviour with dict.values() about pytorch-segmentation HOT 3 CLOSED

Hi, this is surprising, I've never encountered this problem, can you try:

 pixAcc, mIoU, _ = list(self._get_seg_metrics().values())

or to be sure :

seg_metrics = self._get_seg_metrics()
pixAcc = seg_metrics["Pixel_Accuracy"]
mIoU = seg_metrics["Mean_IoU"]

from pytorch-segmentation.

Yeah, I fixed it like the latter. And this one

  for k, v in list(seg_metrics_.items())[:-1]:
    [...]

I replaced with

  for k, v in list(seg_metrics_.items()):
    if k == 'Class_IoU': continue
    [...]

For this:

   pixAcc, mIoU, _ = list(self._get_seg_metrics().values())

Python dictionary doesn't guarantee the order so I doubt this would help.

from pytorch-segmentation.

Its seems that the order of the returned items is the same as the insertion order starting from Python 3.6 (see this PEP )

So I think no need to change the code, if someone else has the same problem, they can see this issue.

Thanks.

from pytorch-segmentation.