API for 2D UNeXt2 model construction needs clarity

[mattersoflight]

The DL@MBL exercise (#100) uses the following call to create a 2D UNeXt2 model.

phase2fluor_model = VSUNet(
    architecture='fcmae', #2D UNeXt2 architecture
...

The above can mislead users into thinking they are working with a masked autoencoder.

@ziw-liu does this also affect #70 ? Can you please fix this while merging #100?

[ziw-liu]

does this also affect #70

I think #70 used a 3D version of UNeXt2.

[edyoshikun]

From the discussion with @ziw-liu, we will move forward by adding temporarily UNeXt2_2D as an alias to fcmae for the 0.2.0 release. Merging the architectures right now is not feasible if we want to load the weights provided in the releases.

After the release, we will merge UNeXt2 and fcmae architectures most likely by deprecating UNeXt2 and adding the missing arguments to fcmae.