Closed tianzhuotao closed 4 years ago
Hi! Thanks for your attention!
Yes. It is a bug. Our code of DeepLabV2 is adapted from the repository: https://github.com/hfslyc/AdvSemiSeg. This bug is existing in their code (please check here).
In our ECCV work, we have found this problem and fixed it. The results are almost the same. I am sorry that I forgot to fix this bug in the release version.
The updated code and pre-trained models will be uploaded soon.
Thanks for your reply~
Hi, the bug is fixed and the new pre-trained models are available now!
Thanks for your great work for tackling the pixel-wise semi-supervised tasks. I am currently following it and I have the following question.
Should the returned value of 'out' at https://github.com/ZHKKKe/PixelSSL/blob/master/task/sseg/module/deeplab_v2.py#L85 be out of the for loop? Otherwise, the ASPP only adds the outputs of dilation rates 6 and 12.
Thanks in advance : )