First of all congrats on the docs, super useful.
Just one small mislead I caught:
In the distance metrics page, for euclidean it states:
- This is used to calculate the distance between two data points in a plane. It is one of the most commonly used distance metric. For an example, see our [image similarity search example].
And in that section of the doc, it uses cosine similarity:
In this example, to compare embeddings, we will use the [cosine similarity score]because this model generates un-normalized probability vectors. While this calculation is trivial when comparing two vectors, it will take quite a long time when needing to compare a query vector against millions or billions of vectors and determine those most similar with the query vector.