On the first run this will take very long because it has to generate the embeddings for 10k documents locally.
First you should import either the plain items or import the items and the embeddings (faster).
Loading...
Then you can run a vector search operation