text-embeddings-inference icon indicating copy to clipboard operation
text-embeddings-inference copied to clipboard

Fix the weight name in GTEClassificationHead

Open kozistr opened this issue 5 months ago • 0 comments

What does this PR do?

Fixes #605

The pooler layer loads its weight using an incorrect key name, causing the classifier and reranker based on GTE to produce wrong outputs.

changelog

  • [x] load the weight from new.pooler.dense.* or pooler.dense.*
  • [x] add the tests
  • [x] update the flash gte test

tests

I've checked that the outputs are aligned with the HF models.

  • [x] WebOrganizer/FormatClassifier
  • [x] Alibaba-NLP/gte-multilingual-reranker-base

Before submitting

  • [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • [x] Did you read the contributor guideline, Pull Request section?
  • [x] Was this discussed/approved via a Github issue or the forum? Please add a link to it if that's the case.
  • [x] Did you make sure to update the documentation with your changes? Here are the documentation guidelines, and here are tips on formatting docstrings.
  • [x] Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag members/contributors who may be interested in your PR.

@Narsil @alvarobartt

kozistr avatar May 03 '25 15:05 kozistr