You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
★★★ Earlier ensure extract result quality instead of letting it enter embedding pipeline which occur error interrupt due to the super high cost of extract
The text was updated successfully, but these errors were encountered:
Hello, what you suggest is perfectly reasonable, however we are relying on pydantic to validate the output of the llm and i'm not sure if it is possible to add a rule like "keep this output if there's at most one error". Something we could do is cache the output for x minutes only if it is a valid output, so when re running the code there would be no need to re evaluate all the correctly sampled chunks. What do you think?
★★★ Earlier ensure extract result quality instead of letting it enter embedding pipeline which occur error interrupt due to the super high cost of extract
The text was updated successfully, but these errors were encountered: