fix: text and character limits on Cohere embedding API #376
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
jimfingal
changed the title
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Outdated suggestions cannot be applied.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.
In a previous PR, I added support for retrieving more than one embedding at once from the Cohere embedding API.
@romenlee pointed out that this causes problems with a large number of texts. It looks like the Cohere embedding API has a limit of either 96 texts, or 2048 characters.
This PR implements more intelligent batching that respects these limits -- it batches texts in chunks that are at most 96 texts or 2048 characters, submits them to the embedding API, and combines the results.
There is a complexity vs. efficiency tradeoff call to make here for the maintainers. This code is much more complex than the original naive code, pre #350 -- but it is much more efficient with network calls so will execute faster. An alternative if we don't want to take on this complexity would be to just revert #350, which would restore functionality for embedding large texts at the cost of giving up performance improvements.