Hi there, There is repo below which helps to solve diarization issues without using clustering, https://github.com/google/uis-rnn Is this right time to do that as an additional feature. Say i have sample dataset of speaker1 and speaker2.