what techniques do you use in Inference API widget?
#63
by
amkazan
- opened
This comment has been hidden
Inference API widget on main model page gives results exactly as I think they should be, when i test my data in it. but after i tried to reproduce it in Qdrant and Milvus dbs I got different (also from each other) not so good results.
what is the process "behind the scenes" of that API widget that does Sentence Similarity? can you please share what is going on so I can reproduce this exact result?
oh! examples of similarity calculation were in description!
amkazan
changed discussion status to
closed