Use some list comprehensions, fall back from multi-chunk encoding #7
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Outdated suggestions cannot be applied.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.
For your delectation, but not necessarily for inclusion -- this change populates
documents
,ids
, andembeddings
with list comprehensions rather than in a loop. (When I tried formingmetadatas
this way, it didn't work, though I think it could be made to.) In my limited experiments, this is somewhat more than 3x faster than before, but I bet that could depend onVECTOR_SEARCH_SENTENCE_TRANSFORMER_MODEL
("BAAI/bge-m3"
in this case).Because the original code was super slow with(Now see the comment below about multi-chunk mode.)BAAI/bge-m3
on 5-chunk pages, I also made the change to give one chunk at a time toembedding_model.encode()
-- this may not be a good idea for other embedding models or other hardware setups.I'm not making any real claims about the performance of list comprehensions here, btw. I started making these changes for aesthetic reasons -- I think they look nice.