Open tomusher opened 2 months ago
The PgVector backend uses its own model where embeddings are stored for indexing by pgvector.
pgvector
It seems in some cases these models can get out of sync with the source Embedding models:
Embedding
>>> PgvectorEmbedding.objects.all().count() 850 >>> Embedding.objects.all().count() 887
Cause to be investigated
The PgVector backend uses its own model where embeddings are stored for indexing by
pgvector
.It seems in some cases these models can get out of sync with the source
Embedding
models:Cause to be investigated