Adding Support for PaliGemma-3B (or rather any other model)

[srinikrish22]

First of all thanks a lot for all the support with making our Jetson devices an AI powerhouse.

I was wondering if there is a possibility to add support to new models or model family. I was thinking about a rough guideline so that the support for new model Architectures can be added (or at least be attempted) to add support via NanoLLM. It is kind of clear that NanoLLM is one of the fastest possible acceleration frameworks for the Jetson devices. Is there at least an outline to get started with added support for new architectures please? I would like to try out some new families of VLM models on my Jetson device, stating with PaliGemma.

Thanks.