Allow batched datasets

Feature request

In Uncle Val I use a specific pipeline, where my model is applied to a single observation, but the loss function may be applied to the whole light curve (or a batch of light curves) only. It would be very convenient if my dataset were allowed to stream data in batches, so I can easily control, that each training batch corresponds to a single (or a few) light curves.

Before submitting
Please check the following:

• I have described the purpose of the suggested change, specifying what I need the enhancement to accomplish, i.e. what problem it solves.
• I have included any relevant links, screenshots, environment information, and data relevant to implementing the requested feature, as well as pseudocode for how I want to access the new functionality.
• If I have ideas for how the new feature could be implemented, I have provided explanations and/or pseudocode and/or task lists for the steps.

The issue names no files, tests, or entry points. Start by locating the dataset and training-batch interfaces that govern streaming, then clarify the expected batching API for single or multiple light curves. Done should include a defined way to yield batches and tests covering the requested training behavior.