[fix
] Ensure that the embeddings from hard negative mining are normalized
#2944
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Outdated suggestions cannot be applied.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.
Hello!
Pull Request overview
Details
If you use
use_faiss
inmine_hard_negatives
, then FAISS expects normalized embeddings. However, the models aren't necessarily normalized. This results innan
for larger datasets and really largescores
for smaller datasets. This PR should resolve that by enforcing that all embeddings are normalized inmine_hard_negatives
.