Propagate kernel size through attention Attention-UNet

[Pkaps25]

Fixes #7726.

Description

Passes the kernel_size parameter to ConvBlocks within Attention UNet, creating a net with the expected number of parameters.

Using the example in #7726 on this branch:

from monai.networks.nets import AttentionUnet

model = AttentionUnet(
        spatial_dims = 2,
        in_channels = 1,
        out_channels = 1,
        channels = (2, 4, 8, 16),
        strides = (2,2,2),
        kernel_size = 5,
        up_kernel_size = 5
)

outputs the expected values:

Total params: 18,846
Trainable params: 18,846
Non-trainable params: 0
Total mult-adds (M): 0.37

Types of changes

• Non-breaking change (fix or new feature that would not break existing functionality).
• Breaking change (fix or new feature that would cause existing functionality to change).
• New tests added to cover the changes.
• Integration tests passed locally by running ./runtests.sh -f -u --net --coverage.
• Quick tests passed locally by running ./runtests.sh --quick --unittests --disttests.
• In-line docstrings updated.
• Documentation updated, tested make html command in the docs/ folder.

[Pkaps25]

@KumoLiu ready for review

[ericspod]

Hi @Pkaps25 I see that this only changes the internal code by propagating the kernel size value, so this should not affect existing code or saved weights since the default everywhere was 3 and still is. Could you please add some tests which create instances of the network with different kernel values and checks the number of parameters changes? We should ensure this behaviour is captured in the tests which, since they pass, all use only the default value of 3. Thanks!

[Pkaps25]

Hi @Pkaps25 I see that this only changes the internal code by propagating the kernel size value, so this should not affect existing code or saved weights since the default everywhere was 3 and still is. Could you please add some tests which create instances of the network with different kernel values and checks the number of parameters changes? We should ensure this behaviour is captured in the tests which, since they pass, all use only the default value of 3. Thanks!

Done!

[KumoLiu]

Thanks for the PR, overall looks good to me.

[KumoLiu]

/build

[KumoLiu]

/build

[KumoLiu]

/build