Repository navigation
Conversation
CatOn60Hz
marked this pull request as ready for review
October 17, 2026 08:35
KennethEnevoldsen
left a comment
KennethEnevoldsen
left a comment
Contributor
There was a problem hiding this comment.
generally looks good
Author
|
@KennethEnevoldsen can u run the test now ?? |
Samoed
reviewed
Oct 21, 2026
The 3,956-line benchmarks.py is split into modality/domain modules (text_language, text_multilingual, code, retrieval, image, audio_video, domain) plus _citations.py for shared citation constants. __init__.py re-exports the identical public set, so the namespace-scan registry in get_benchmark is unchanged. Verified against main: registry names (78), aliases (42) and per-benchmark task sets are all identical; ruff and the benchmark test suite pass. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Address review feedback: drop _citations.py and move the citation constants to where they are used. _MAEB_CITATION is only used by the audio benchmarks, so it lives in audio_video.py; MMTEB_CITATION is defined in text_language.py (its primary users) and imported by the code and multilingual modules. Registry names (78), aliases (42), task sets and citations remain identical to main. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
CatOn60Hz
force-pushed
the
refactor/split-benchmarks-module
branch
from
October 21, 2026 09:54
d6bb023 to
5cfa9d6
Compare
This branch has not been deployed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Splits the 3,956-line
mteb/benchmarks/benchmarks/benchmarks.pyinto themed submodules by modality/domain:text_language.pytext_multilingual.pycode.pyretrieval.pyimage.pyaudio_video.pydomain.py_citations.pyMMTEB_CITATION,_MAEB_CITATION__init__.pyre-exports the identical public set, so the namespace-scan registry inget_benchmarkis untouched.rteb_benchmarks.pyis unchanged.Verification
Compared the live registry against
main:ruff format/checkclean; benchmark test suite: 643 passed, 71 skippedNotes for reviewers
__init__.py(and thus unregistered) onmain—MTEB_AFRICA,MTEB_AFRICA_LITE,IFIR,BEIR_PL,VISRAG_RETRIEVAL.VISRAG_RETRIEVALlooks intentional-but-forgotten; happy to export it in a follow-up if desired._This was done with the help of AI and verified by me