feat: Add model kwargs to SentenceTransformersRanker#6627
Merged
Conversation
sjrl
requested review from
dfokina and
vblagoje
and removed request for
a team
github-actions
Bot
added
topic:tests
topic:build/distribution
topic:CI
2.x
type:documentation
vblagoje
reviewed
Dec 22, 2023
vblagoje
approved these changes
Dec 22, 2023
Contributor
Author
|
Hey @vblagoje I made the change as we discussed. Let me know what you think! |
Member
|
This version is better; it is much more explicit about what is happening. Thank you @sjrl |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Related Issues
Proposed Changes:
Updates cross encoder init to take model kwargs so we can pass float16 as a loading option. Increases speed of ranking 40 docs by 4x when using float16.
How did you test it?
Notes for the reviewer
Checklist
fix:,feat:,build:,chore:,ci:,docs:,style:,refactor:,perf:,test:.