WIP: Update outlines (and outlines prompts) and transformers versions

[dmjoy]

@eveenhuis @jadie1 @barry-ravichandran FYSA

This is a WIP, but wanted to draw attention to some of the changes that need to be made for the latest outlines version. Specifically:

• The outlines.pompt decorate has been removed, now the guidance is to use the outlines.Template.from_string function.
  • Copied this decorator in from the older outlines version for near term "compatiblity" rather than require changes to all of these old templates (using outlines.Template.from_string we also need to change how these templates are invoked wherever they're used, and that's scattered around several algorithms; not worth) future templates should strongly considering using this new approach though
• The notion of a sampler has been removed from outlines, instead inference control parameters are just passed through to whatever backend is loading / serving the model (so for determinstic sampling using transformers as the model provider we would set do_sample=False in the generation_kwargs)

[dmjoy]

Alright so functionally this all seems to work, but when running the integration tests, there are several differences (mostly that makes sense as enough pieces have changed here that I wouldn't expect to get the exact same output). However, it seems like there's some non-determinism even with do_sample=False, and for the comparative regression ADMs I'm getting negative regression values in some cases, which I don't think we've seen before. Digging into the negative regression values though and it seems like maybe this was never enforced with JSON schemas?
dottxt-ai/outlines#215
dottxt-ai/outlines-core#147

I need to spend some more time with this to understand what's really going on before feeling confident in merging. Anecdotally it does seem a bit faster now, but if it's not deterministic and doesn't enforce the entire schema what's the point.

[barry-ravichandran]

@dmjoy Tagging this package update PR here in case it's ok with you to merge it into your PR to fix some issues with installing via poetry: #250

[alright-code]

The problem causing the non-determinsm and invalid regression values was improper handling on the outlines side of vocabulary creation: dottxt-ai/outlines#1831. This was merged recently into outlines main, but we are blocked from directly using this due to vllm dependencies. I've implemented a monkey patch fix we can use for the time being.

There was a second bug in outlines here: dottxt-ai/outlines#1817. Same thing with above, we are blocked from the official fix because of vllm.

[dmjoy]

@alright-code I think these changes on the whole are fine, I just left a few comments / change requests. I believe the integration test differences are within reason, but before merging I will want to check out this branch and run one or two tests with the latest data and configs to make sure the scores we're getting back are still good.

[dmjoy]

There's already some mechanism(s) for filtering out lines during the integration test diff. We probably want to exclude these "HTTP Request" lines as they can be non-deterministic

[dmjoy]

Seems like you still might need to re-generate these test files (I see you added the HTTP filter but these lines are still in the output)