`tf.keras` version that allows any input resolution and doesn't use `Lambda` layers

[prhbrt]

Since the UNET architecture only uses layers that can scale with the image dimensions, the fixed dimensions seem artificial. I've added a zero-padding layer that increases the dimensions to the nearest multiple of 32. The padding is cut off in the end.

Moreover, I've removed the lambda-layer, as it creates marshaling warning, and used ZeroPadding2D's asymmetric padding feature. This removes a warning upon load_model.

I converted the eynollah-models to this architecture, and they should load without warnings now and use the tensorflow.keras API and can be found here.

This might allow you to skip the patching as used in Eynollah and speed up the whole process. Please let me know what you think.

Notes and sanity checks:

• the enhancement model uses the light version without the last batchnorm and with a sigmoid rather than softmax.
• the column classifier is a different architecture, I just copied the original model, and hence it's not dimension invariant.
• Two models weren't in savemodel-format on your website, e.g. the light versions eynollah-main-regions_20220314 and eynollah-textline_light_20210425, so I couldn't convert them.

[cneud]

Hi @prhbrt, thanks a lot for looking into this and for contributing!

FYI, we are planning to update and refactor this repo and integrate the model training code with https://github.com/qurator-spk/eynollah for future maintenance, so this comes in very handy.

My colleagues @vahidrezanezhad and @michalbubula will be working on this - although due to various reasons, we likely won't be able to get our hands dirty much before March. But we will try our best to review and merge any contributions also beforehand.

Btw the two models that you were missing should be available from our HF:

• eynollah-main-regions_20220314 -> https://huggingface.co/SBB/eynollah-main-regions
• eynollah-textline_light_20210425 -> https://huggingface.co/SBB/eynollah-textline_light

[prhbrt]

@cneud Understood! Could you in the meantime provide a list of all model-architectures used for eynollah (specifically python code)? I couldn't find python-code for the column classifier in particular, so that one still has fixed dimensions.

Also note that this yolo-version might give slightly different outputs as your patching example, due to boundary conditions.

Thank you in advance!

[vahidrezanezhad]

@cneud Understood! Could you in the meantime provide a list of all model-architectures used for eynollah (specifically python code)? I couldn't find python-code for the column classifier in particular, so that one still has fixed dimensions.

Also note that this yolo-version might give slightly different outputs as your patching example, due to boundary conditions.

Thank you in advance!

@prhbrt
Sure, I'll try to make classifier code public in the meantime.