Find answers from the community

Updated 2 years ago

Token usage

Ok I’ll check that out. And interms of calculating tokens manually what components contribute to the total token usage?

Just thinking of doing a manual calculation to understand ahah

1 comment

LLogan M

The embedding model and llm predictor are the two root elements that contribute to token counts

Internally, there is a function decorator that is used to count tokens for different functions in different places

Add a reply