Provide an example model in the getting started guide

[GregoryComer]

📚 The doc issue

With the new getting started guide, we should provide a model in the guide, so that users can follow along without having a specific model in mind. We ideally want to balance simplicity and it being interesting. Perhaps one of the smaller torchvision image classification models? I'd like to be able to keep the code snippets completely inline and with minimal dependencies, so it might be tricky to balance the extra code for image handling.

Suggest a potential alternative/fix

No response

cc @mergennachin @byjlw

[byjlw]

what about resnet-50?
Also to be clear, we're just downloading in the step right? We won't store a copy in the repo

[GregoryComer]

what about resnet-50? Also to be clear, we're just downloading in the step right? We won't store a copy in the repo

Yeah, I was thinking either MV3 or ResNet. Seems like a good choice. On the PTE, I definitely agree on not putting in tree. S3 or HF or whatever should be good. We probably just want to be able to programmatically regenerate it, maybe as part of the docs build job.

[mergennachin]

#8179