Improvements for user interface and tutorial

This release contains several improvements to user interface. There are some changes to improve how cellpose is integrated and allows other pytorch's DataLoader-based training to be implemented.

Additionally, a full tutorial has been added to the documentation on how to fine tune a cellpose model. And an example of how to register a micro-sam-based segmentation method has been made available in the notebooks folder.

The package dependencies have also been updated to restrict the use of zarr v2, this until other dependencies are updated to use zarr v3.