Collected sources and patterns will appear here. Add from search, explore, or the patterns library.
(List<WordWithTimestamp>, List<DiarizationSegment>) -> List<DiarizedWord>
Map speaker labels from an independent diarization timeline to transcribed words by finding the maximum temporal overlap between word segments and speaker segments.
Problem it solves
ASR pipelines and speaker diarization pipelines output non-aligned, independent sets of time intervals.
Consumes
Emits
The real projects this mechanism was found in. Attribution is the point — this is how the best teams actually do it.