Vidit

Thanks for the help.

Hi, the shape of Q is (B, H*W, V) and that of Z is (B, d, H*W, V). So when torch.mul(Z,Q) is used, output shape will be (B,d, HW, V)....

Regarding `torch.mul(z,Q)` I realized it will throw an error if batch size if greater than 1. Thanks for pointing it out. You can reshape Q as (B, 1, HW, V)...

Another small thing, in file `graphLayer.py` in [line 193](https://github.com/vidit98/graphconv/blob/fa5b72774361a52173e0c55598c7f4e379dd5844/graphLayer.py#L193) you need to change `Z`(capital Z) to `z`(small z) in `torch.div(Z, norm)`.

Hi, thank you for your interest in the work. You can find the model [here](http://sceneparsing.csail.mit.edu/model/pytorch/ade20k-resnet50dilated-ppm_deepsup/).

Yes, I also faced this issue once, it is generally border case which causes problem for example numbering in ground truth starts from 0 and you have started it from...

There can be a possibility that the dataset has changed, you can check whether they have increased the number of classes from 150. Another thing you can check is that...

This would be a cool feature to add. If we can initialize mask in `sketch` using a predefined mask and ask user to refine it, would be really useful