Clemens Neudecker

Let me first confirm the above and then we can rename the option, ideally also consistent for scaling, enhancing, resizing.

@M3ssman Python 3.8-3.11 should all be working/supported - problems with compatibility start with >=3.12 ;)

> LGTM, much improved docs, many thanks. > > But can we keep the PEP8 changes out of it for now? They will be cleaned up later but there's now...

> I refactored the sbb_pixelwise_segmentation training sources to pythorch and is damn faster than the tensorflow version and easy to port. @johnlockejrr do you happen to have changes related to...

> https://github.com/johnlockejrr/sbb_pytorch/blob/main/CHANGELOG.md > https://github.com/johnlockejrr/sbb_pytorch/blob/main/TECHNICAL_CHANGELOG.md Thanks you! Unfortunately the repo is not accessible yet, maybe it is set to private?

Very interesting, thanks! TBH the main reason preventing us from upgrading is the switch from Keras v2 to Keras v3 API since TF 2.13. Apart from upgrading the codebase to...

We already use PyTorch in some other projects, also [timm](https://github.com/huggingface/pytorch-image-models) lures. Switching could provide an opportunity to revisit GPU code and make it more efficient. Certainly an option to consider,...