Backoff

[shaankhosla]

Is there a way to configure the retry parameters for any of the response completion requests? For example, if I hit a rate limit, can I have it be set to retry 3 times with exponential decay?

[shaankhosla]

Alternatively, can I remove all the retry logic within async-openai so that I can wrap my API call with https://docs.rs/backon/latest/backon/

[64bit]

Because it would be a breaking change, can you elaborate on how the new Client API with backon will look like?