0.3

Add the ability to set the max_tokens_to_sample param. This is described as The maximum number of tokens to generate before stopping. If not set, this defaults to a generous 10_000 tokens. #4