Allow max_tokens = -1 for ChatOpenAI

[adlindenberg]

The Chat API allows for not passing a max_tokens param and it's supported for other LLMs in langchain by passing -1 as the value. Could you extend support to the ChatOpenAI model? Something like the image seems to work?

[Aratako]

This solution seems to effectively resolve the issue.
Do you plan to create a pull request yourself? If not, I can make the necessary changes. Would that be alright?

[adlindenberg]

@Aratako that'd be fine with me, thank you!