Better auto batching for resolve LLM calls #128
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Staging to main
feat: add reduce operation lineage
fix: change gleaning prompt to validation_prompt
This reverts commit bcd4e8c.
…item-renaming ucbepic#91 document > item renaming
docs: add 'output' argument to ResolveOp code eg
fix: edit agent for synth resolve task
docs: add sample and cluster to docs
fix: make docs work
feat: adding human in the loop for split-map-gather decomp
fix: cache partial pipeline runs
mark as flaky test
only compare distinct pairs in resolve
LinkResolveOperation
Fix Resolve and Map progress bars
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Making some changes to auto_batch in resolve to heuristically understand the workload and decide batch sizes. Also implemented a way to fit batches that don't redo previously clustered operations, and therefore more efficiently use parallelization for llm calls.
Testing:
See test_resolve_auto_batch, for a sample test suite to run.
Not Done:
Model limits are hardcoded to 500 requests concurrently at the moment. We need to find a better way to incorporate this into the model api.