the kernel

[maxmit233]

Hi, author， There was a variable in the code called diagnoal_weight_filter I think the variable should change in the trainning time,but it never changed when I debugging. It's so confusing. And I wonder if the variable conduct the same role as the diagnoal_weight_filer in the tensorflow implementation will change.

[benedekrozemberczki]

Do you have some output for this?

On Sat, 16 Jan 2021 at 12:49, maxmit233 notifications@github.com wrote:

Hi, author， There was a variable in the code called diagnoal_weight_filter [image: 屏幕截图 2021-01-16 204442] https://user-images.githubusercontent.com/29227614/104812208-b05afd80-583b-11eb-82d4-5908c1c59b00.png I think the variable should change in the trainning time,but it never changed when I debugging. It's so confusing. And I wonder if the variable conduct the same role as the diagnoal_weight_filer in the tensorflow implementation will change.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/benedekrozemberczki/GraphWaveletNeuralNetwork/issues/9, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEETMF2V3K3WI27BLJABGFTS2GDMJANCNFSM4WFHV3JA .

[maxmit233]

Do you have some output for this? … On Sat, 16 Jan 2021 at 12:49, maxmit233 @.***> wrote: Hi, author， There was a variable in the code called diagnoal_weight_filter [image: 屏幕截图 2021-01-16 204442] https://user-images.githubusercontent.com/29227614/104812208-b05afd80-583b-11eb-82d4-5908c1c59b00.png I think the variable should change in the trainning time,but it never changed when I debugging. It's so confusing. And I wonder if the variable conduct the same role as the diagnoal_weight_filer in the tensorflow implementation will change. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#9>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEETMF2V3K3WI27BLJABGFTS2GDMJANCNFSM4WFHV3JA .

As you can see, after I call the backward() method of the loss .the grad attribute of the diagnoal_weight_filter is None. It is unable to change.

[benedekrozemberczki]

This seems to be a pytorch geometric error.