nabila
nabila
hi all -- any update on this?
Did you guys split up the training masks to one hit encode them? I have a memory error so I'm wondering what everyone else is doing for it
This is a good question! I wonder if it has to do with weight sharing? It says in the paper that "each capsule in the [6 × 6] grid is...
Is there a suggestion on a potential solution I could try to implement? At the moment, I have a Bert base +2 adapters saved as two seperate onnx models resulting...
@hadware did you figure out if it does?
Hi Jacob! I will upload it as soon as I am free but it will be a BSD license! Thanks for reminding me :)
Hi John, I initialize all weights to follow `glorot_normal ` distribution but did not experiment with the original paper's init method. Is there a large difference in DSC when using...
Hi, thanks for the interest in our work. Since this is segmentation, the "CAM" can be thought of as the feature map activations at any layer. Since in gradCAM and...
Hi jie, I appreciate your interest in our work. I would like to ask that you limit your issues to one thread because you have re-opened issues that are not...
Hi Kevin, can you check the range of your predictions and the range of your ground truth? Both masks should be in range [0-1]. You can just print the max...