[PR #11881] fix(ml): tokenization for webli models #13398

Closed
opened 2026-02-05 15:14:01 +03:00 by OVERLORD · 0 comments
Owner

Original Pull Request: https://github.com/immich-app/immich/pull/11881

State: closed
Merged: Yes


Description

The new round of models assume the text is canonicalized. This PR checks the model config file to see if this needs to be done for a model. If so, it removes punctuation and converts the text to lower case.

Tested that search results are relevant regardless of casing.

Fixes #11865

**Original Pull Request:** https://github.com/immich-app/immich/pull/11881 **State:** closed **Merged:** Yes --- ### Description The new round of models assume the text is canonicalized. This PR checks the model config file to see if this needs to be done for a model. If so, it removes punctuation and converts the text to lower case. Tested that search results are relevant regardless of casing. Fixes #11865
OVERLORD added the pull-request label 2026-02-05 15:14:01 +03:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: immich-app/immich#13398