Using OpenAI embeddings on Supabase database - SupabaseTips

preview_player
Показать описание
Supabase has built-in support for pgvector, which can be used to store embeddings generated on OpenAI. This allows developers to build advanced search engines for text documents on top of OpenAI.

In this SupabaseTips video, we look at what embeddings are, and how you can build your search engine using OpenAI and Supabase.

More Supabase 👇

Рекомендации по теме
Комментарии
Автор

this kind of content is much useful than a hour long boring stuff. nice edit too

ianjaspersantos
Автор

At 0:49 `vector` is added to the `extensions` schema. When I did this I was unable to add a column of type `vector` to my table from the Table Editor UI. I needed to add `vector` to the `public` schema for this to get things working. Not sure if this is a limitation of the editor or if this video was based on a recording that was later changed.

daveteare
Автор

Please do a more detailed video on this subject please. I was able to follow up until 1:02 then you showed some partially hidden screen with some JS script in it. Is this run in SupaBase, can it be run in SupaBase? You do the same thing again at 2:37. Can you do a video please for newbies on the sections that you partially hid, I and others would appreciate that.

AssassinUK
Автор

This is such an amazing video, it was concise and was well explained, thank you so much!

How can accomplish this using open-source embeddings such as HuggingFace Embeddings? Is this possible?

luisxd
Автор

Question: the function that is mentioned in the blog is always referencing the same table. Can I make the table also as an argument of that function?

dejoma.
Автор

How can we delete embeddings using metadata filters ?

louislecouturier
join shbcf.ru