Dr. Bruce Hashemian
Dr. Bruce Hashemian
Hi @Nic-Ma @wyli @rijobro @ericspod @yiheng-wang-nv, Do you have any insight why torchscript is failing here? https://github.com/Project-MONAI/MONAI/actions/runs/3190912645/jobs/5206609165#step:7:30985 It seems that it is related to SobelGradients but do you know what's...
> > As your error message shows: _Module 'HoVerNetLoss' has no attribute 'sobel' (This attribute exists on the Python module, but we failed to convert Python type: 'monai.transforms.post.array.SobelGradients' to a...
I try to resolve some failing tests and some formatting, and then it should be ready. I'll request to review afterwards.
Hi @wyli, sure, I'll look into it.
@myron do you have any insight here?
It seems that this is a PyTorch issue that is caused by using MetaTensor (a subclass of `torch.Tesnor`). @wyli has created a bug report on PyTorch: https://github.com/pytorch/pytorch/issues/86456
@JHancox do you have any comments on this PR?
@cooperlab what do you think?
Hi @KumoLiu @JHancox, As @wyli mentioned, Sobel gradients in OpenCV for higher dimensions is implemented differently but they should match exactly for a kernel size of 3x3. The main reason...
Hi @KumoLiu, thank you very much for comparing the results. It seems that it is related to threshold but in fact it shouldn't. The only thing that I can think...