cvqluu/simple_diarizer

Combines Silero VAD for speech detection, SpeechBrain embeddings (X-Vector or ECAPA-TDNN) for speaker representation, and spectral or agglomerative clustering to identify and separate speakers. Optionally integrates ESPnet ASR models for transcription alongside diarization. Supports configurable embedding and clustering methods, allowing users to trade off between speed and accuracy for different use cases.

155 stars. No commits in the last 6 months.