Mteb

Latest version: v1.34.14

Safety actively analyzes 706267 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 32 of 75

1.12.81

Fix

* fix: Add bm25s as a model for retrieval (1082)

* add bm25s as a model
* getting bm25s working
* add nf corpus results
* add as optional dependecies
* add bm25 to make install-for-tests
* move results to results repo
* add run script
* make test passing
* make lint
* types

Co-authored by: malteos <githubi.mieo.de> ([`5269f2c`](https://github.com/embeddings-benchmark/mteb/commit/5269f2c4538cc3b0e691de235d1cd2011e1bab5a))

1.12.80

Ci

* ci: fix ci docs to work on prs and in main (1084)

* ci: allow docs ci to be run by contributors on PRs

* ci: Add missing &34;fi&34; at the end ([`c977d35`](https://github.com/embeddings-benchmark/mteb/commit/c977d35996c12ee94546470bb7875d68bb304084))

* ci: allow docs ci to be run by contributors on PRs (1081) ([`db45092`](https://github.com/embeddings-benchmark/mteb/commit/db45092f31a1be05f264a0752fa68964a35b5927))

* ci: run docs ci on PRs (1078) ([`bf02080`](https://github.com/embeddings-benchmark/mteb/commit/bf02080f6062bc28bff597663ed1759a63233535))

Fix

* fix: ensure that CQADupstackRetrieval is included in results if possible (1079)

* move mmteb script into correct folder

* remove unec file

* add results to gitignore

* fix: ensure that CQADupstackRetrieval is included in results if possible

* remvoe cqadupstack script

* Added fixes based on corrections ([`3109b4d`](https://github.com/embeddings-benchmark/mteb/commit/3109b4deaf2ca080036240d4f734b304da7fc23a))

Unknown

* Add support for the Bright benchmark (971)

* Bright

* Update mteb/tasks/Retrieval/eng/BrightRetrieval.py

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com>

* Update mteb/abstasks/TaskMetadata.py

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com>

* Update mteb/tasks/Retrieval/eng/BrightRetrieval.py

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com>

* Update mteb/tasks/Retrieval/eng/BrightRetrieval.py

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com>

* Update mteb/tasks/Retrieval/eng/BrightRetrieval.py

Co-authored-by: xiamengzhou <296337231qq.com>

* Create run_mteb_bright.py

* Fix format

* Fix splits

* Update metadata

* Rmv newline

---------

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com>
Co-authored-by: xiamengzhou <296337231qq.com>
Co-authored-by: hongjin-su <114016954+hongjin-suusers.noreply.github.com>
Co-authored-by: Niklas Muennighoff <n.muennighoffgmail.com> ([`e05c53c`](https://github.com/embeddings-benchmark/mteb/commit/e05c53cbc7a5be62090de156a033689f16cab040))

1.12.79

Fix

* fix: Added models x tasks which haven&39;t yet been run (1038)

* Added models x tasks which haven&39;t yet been run

* removed duplicate citations

* adapt to new results repo structure

* Rename PawsX to PawsSXPairClassification

* rename PawsX class

* updated missing tasks

* Updated missing tasks

* tmp

* format ([`ae2e2b7`](https://github.com/embeddings-benchmark/mteb/commit/ae2e2b7c8140176f0aa4f3d7d90856f5acbbabb6))

1.12.78

Fix

* fix: Multi-gpu/simpler models; Add gte (1059)

* Merge

* Adapt

* Simplify

* Check for rev again

* Rmv cmmnt

* Simplify

* simplify

* Rmv comment

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com>

* Use logging; change try except; add info

* Lint

* Rmv results

* Update rev

* format

---------

Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsengmail.com> ([`66ae979`](https://github.com/embeddings-benchmark/mteb/commit/66ae979f29eef1d66cb98870e19773ca76d30819))

1.12.77

Fix

* fix: Added benchmark construction (1003)

* Added utility function for downloading models results

* Added bench construction

* format

* Added correction citations

* format ([`dd76c67`](https://github.com/embeddings-benchmark/mteb/commit/dd76c674b86c7370de73a3d4d49f9e161f77ea96))

1.12.76

Documentation

* docs: Update contributor table (1071)

update contributor table ([`778d7a3`](https://github.com/embeddings-benchmark/mteb/commit/778d7a3bf85b2023cc8ba9b2c35a810dcfa5e924))

Fix

* fix: restructure test suite to avoid to many external dependencies (1070)

* fix: restructure test suite to avoid to many external dependencies

The intention behind this is to make tests faster and more robust. Especially to avoid false positives where the test works, but fails due to interaction with HF, the OS (CLI commands).

- Restructure test suite to make it easier to find specific tests and add tests in the future. e.g. added all evaluators test to a folder
- splits test_mteb into multiple segments, one for testing SentenceTransformers, one for testing evaluation workflow, one for testing integration with datasets
- avoid using encoders when testing evaluation workflow (one extra interaction with hf which can fail and it is slow)
- Added MockEncoders for testing torch, numpy and torch bf16 return types (see FP16 embeddings fail with BitextMining 1064)
- Added MockTask for all task types to avoid downloading too many datasets and to allow for more diverse without making the test suite too slow

The slowest test are now the mteb_rerank tests along with the integration tests. orionw was hoping we might be able to to replace the dataset with a dummy dataset. However the slowest atm. is 20s

This PR also added a few minor fixes (surprisingly few actually) for handling torch bf16 which was not handled and a minor fix to AbstaskClusteringFast (isaac-chung)

* fix: Remove precision for more readable scores and lower memory footprint (1075)

* fix: reduce precision of results for readability

* Added example cases

* Removed &34;dev&34; as a default split from flores

* remove flores results

* fixes errors due to merge ([`636018b`](https://github.com/embeddings-benchmark/mteb/commit/636018be6ec04a2051881abd0068ce018e485be5))

Refactor

* refactor: update TaskMetadata (1076)

* update TaskMetadata.py

* update _add_stats.grit

* update _add_stats.grit

* update 415 files

* update _add_stats.grit and TaskMetadata.py

* update _add_stats.grit

* update 86 files

* update 6 files update 341 files

* delete 1 file and update 416 files

* misc. ([`57c1c12`](https://github.com/embeddings-benchmark/mteb/commit/57c1c12727b6300465cb647f2e55c9af0049d759))

Unknown

* Update points table ([`cff1ae6`](https://github.com/embeddings-benchmark/mteb/commit/cff1ae6c2515b8aa97b15da7d72c47b824c7d94f))

* Update points table ([`86eb18c`](https://github.com/embeddings-benchmark/mteb/commit/86eb18c1c8a2d2c04c0a2752fd9302726119100f))

* Change refresh description (1074)

Update adding_a_model.md ([`d98dcd8`](https://github.com/embeddings-benchmark/mteb/commit/d98dcd8958ecaeda9b7f97f340412be42f6f9efb))

Page 32 of 75

Links

Releases

Has known vulnerabilities

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.