[Notebook] request for a multi output multilabel problem.

First of all thank you for this awesome project. I am using this to get started in ML/AI.
The [Zero Shot Notebook ](https://github.com/huggingface/setfit/blob/main/notebooks/zero-shot-classification.ipynb) helped me a lot to understand the first couple of things. I love how you showed different approaches e.g. "true" zero shot, but also with some synthetic examples to get better results.

Unfortunately nearly all tutorials and guides that i have read so far, show a multi label example with exactly one label category.
In the Zero Shot Notebook this would be the `emotions` and its array of labels which are `['sadness', 'joy', 'love', 'anger', 'fear', 'surprise']`

I'd love to see an example notebook with multiple categories.
For example:
`emotions: ['sadness', 'joy', 'love', 'anger', 'fear', 'surprise']`
`intend: ['informative', 'commercial', 'influencial']`
`readability: ['low', 'medium', 'high']`

Is this something that can be done with setfit? Are there different approaches to tackle a problem like this?

A Notebook explaining that would help me and hopefully others.

PS: Sorry if this is an easy or entirely different ML/AI problem, i am just getting started