[Client] Add guided_grammar and other missing fields (#532)

Add guided_grammar to the client, + add some missing fields to some codepaths

Image cache and balloon on H100s, also temporarily stop people from u…

…sing A100 (#523)

* Cache H100

* Stop people from using A100

* no cover

* no cover

* update client version

Fix package name mapping (#514)

Guided decoding (#476)

* Guided decoding

* endpoints

* fix

* update client

* unit tests

* fix test

* coverage

* coverage

* fix

* try to bump coverage

* more tests!

* lint

[Client] Add num_prompt_tokens to the client's CompletionOutputs (#467)

* add prompt token, untested

* comment

* remove stop_str stuff, it doesn't do anything with the public api, and it breaks on certain frameworks when hosted locally

Change back batch infer GPU util and add tool completion client chang…

…es (#465)

* Change back batch infer gpu util

* Add client changes

* fixes

* bump

[Client] Add Auth headers to the python async routes (#434)

* add Auth headers to the async routes

* bump version

Minor fixes for batch inference (#426)

* Fix file not found

* progress fix

* add tests

* bump

* typing

pypi version bump (#391)

publish 0.0.0b19 (#322)