how do you compute the similarity

#4
by G0tto - opened

Does it based on cosine similarity?

Sentence Transformers org

Hello!

It looks like this model actually uses dot product to compute the similarities (source 1, source 2), unlike most models which indeed rely on cosine similarity.
So, the usage in the README looks to be correct.

  • Tom Aarsen

I understand,thanks for answering!

Sign up or log in to comment