Is there any easy way to just get the generated tokens instead of the full output?

[arnavgarg1]

Same as the discussion title - I am wondering if I can just get the generated tokens in the response as opposed to the input tokens + generated tokens.

I couldn't see anything around this here: https://github.com/bentoml/OpenLLM/blob/main/src/openllm/client/runtimes/base.py. Probably just needs a new return_response strategy?

[aarnphm]

This is working in progress, in the same line with SSE