Hans-Jürgen Schönig builds some indexes on vector data:
In the previous post we have imported a fairly large data set containing Wikipedia data, which we downloaded using pgai. However, importing all this data is not enough because we also need to keep an eye on efficiency. Therefore, it is important to understand that indexing is the key to success.
Click through for an overview of what options are available for indexing vectors in PostgreSQL, as well as the trade-offs and disk space requirements.