Mennatullah Siam

Nope. But you can check this [repo](https://github.com/meetshah1995/pytorch-semseg) its very useful that has multiple segmentation architectures but not aimed at real-time performance. You can use it as a start and add...

First you have to use inference mode this will call test_inference method in [train/train.py](https://github.com/MSiam/TFSegmentation/blob/master/train/train.py). Second the most efficient architecture we used is SkipNet ShuffleNet not UNet that provides 143 fps...

In order to run the inference mode. Just uncomment line 919, which is setting is_training placeholder to false. It was commented because I removed this variable when I was doing...

Thats still very slow on 1080 Ti! Probably you're using a higher image resolution than what we were measuring with. Regardless please refer to this issue for further help on...

Thanks. I have all experiments described here https://github.com/MSiam/AdaptiveMaskedProxies/blob/master/Experiments.md

I also wasn't sure what to use in the weightedBCE code for the threshold as in the paper you use 0.2 but the code seems to just use anything above...

I also had a question on your PyramidDilatedConv module in line https://github.com/tfzhou/MATNet/blob/c8b95e527c486c304f711cc7dffb060f31abe19f/modules/MATNet.py#L446 it seems to only use block1 and block2, block3 are not used. Is this an issue? Thanks for...

I am facing the same error actually. It turned out to be because the val_loader was empty it wasn't able to load any images. If you change this line in...

Thanks. The dataset will be released upon acceptance so the paper is still under review sorry about that. However the details of the contrastive learning method is available here: https://arxiv.org/abs/2105.03533

Thanks for your response. I actually tried what was reported in the issue I linked to in my post: ``` "Please provide the bounding box coordinate of the region this...