for this and it works um and it returns embeddings that are stored in qdrant by brai

which then attempts to match them i wonder if i just need to make give him authority to

loosen the requirement so that the matches can be more easily made automatically across phone

calls and transcriptions as opposed to currently where it's asking me to identify them all

ways because it's very rarely actually finding a match maybe one out of 30 is getting a match

one out of 30 utterances is getting a match currently but i also see that the speaker

labels segments that are detected by pioneer aren't always the same segments that return by

whisper so i don't know if there's some mismatch there i got confused about how it actually works

i'm also interested in possibly

you