Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What this PR does / why we need it:
This PR introduces the cascade functionality to Libem, enabling the system to dynamically adjust model usage based on confidence scores. This aims to optimize cost and throughput while maintaining the accuracy of GPT-4o.
Key changes include:
vectorize
,prematch
, andmatch
under Libem/cascade.prematch
andmatch
stages.Libem/examples/cascade
:util.py
: Contains benchmark and plot functions to measure cost and accuracy.online.py
: Provides a working example of the cascade functionality.bench.py
: Performs benchmarking and generates graphs to visualize cost and accuracy.Which issue(s) this PR addresses (if any):
Addresses # NA
Special notes for your reviewer:
I am only adding the reset as part of the cascade since it is the most reasonable place I can think of for now. Would love to improve on this later if there is a better place to do reset.
Does this PR introduce a user-facing change?