Haystack on NVidia Jetson

[Solido]

Hi!

I'd like to run Haystack on edge computing and I just starting to assess feasibility.

The idea would target NVidia Jetson platform to benefits from ML acceleration.
Is there any track record on this already?

Here are some early questions that can help me moving forward and would really appreciate feedbacks and suggestions:

• Is there any feedbacks or requirements to run it on a middle sized QA system?
• Is there a specific Jetson hardware that you would suggest for acceptable perfs? (Nano 2Go or more?)
• Would you suggest specific datastore like Milvus or SQLite for embedded perfs?

Thank you Haystack's community!