An example or recommendations for creating large image datasets?

I know that Apache Beam and `datasets` have [some connector utilities](https://huggingface.co/docs/datasets/beam). But it's a little unclear what we mean by "But if you want to run your own Beam pipeline with Dataflow, here is how:". What does that pipeline do?

As a user, I was wondering if we have this support for creating large image datasets. If so, we should mention that [here](https://huggingface.co/docs/datasets/image_dataset).

Cc @lhoestq