feat(search): dense semantic search with v5 embedding model#85
Merged
Conversation
kdroidFilter
force-pushed
the
feat/v5-dense-search
branch
8 times, most recently
from
2cd5766 to
b9e938b
Compare
- Add SeforimEmbedder (ONNX v5 int8) + HebrewV5Normalizer (final-letter folding) - HybridSearchEngine (BM25 + dense, RRF) + VectorSearcher over a fused Lucene index - BuildVectorIndex / fused KnnFloatVectorField indexing in the generator - Bundle + fetch the v5 model (PackageArtifacts, DownloadEmbedModel -> v5-int8) - CI: free disk space on the runner before the build
kdroidFilter
force-pushed
the
feat/v5-dense-search
branch
from
b9e938b to
8f8da4e
Compare
Contributor
|
מתוכנן לשחרור ציבורי? |
Owner
Author
|
אני לא יודע, אוליי, אבל מה זה יעזור לכם ? |
Contributor
|
זו כבר בעייה שלנו 😉 |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Adds hybrid lexical + dense semantic search powered by the v5 Hebrew/Aramaic embedding model.
SeforimEmbedder— ONNX (v5 int8) query embedder;HebrewV5Normalizer(strip nikud/teamim + final-letter folding) to match training.HybridSearchEngine— fuses BM25 + dense (RRF);VectorSearcherover a fused Lucene index (KnnFloatVectorField, cosine).BuildVectorIndex+ fused text/vector indexing.PackageArtifacts);DownloadEmbedModelfetches thev5-int8release.Model
v5b (full-corpus, final-folded). Fair link eval on identical 2000 pairs: R@1 0.461 / R@10 0.870 / MRR 0.606 vs v4 0.418. int8 quantization ≈ 0 quality loss (cosine 0.999).
Notes