Model reference · open weights

speaker-diarization-3.0

speaker-diarization-3.0 is an open-weight audio or speech model from pyannote.

For self-hosting, compare this build's requirements below with the GPU you plan to rent.

• speaker-diarization-3.0 is an open-source pipeline by pyannote designed for automatic speaker diarization.
• It processes mono audio sampled at 16kHz and outputs speaker labels, automatically handling resampling and downmixing of multi-channel inputs.
• The model is trained on datasets including AISHELL, AMI, and VoxConverse, and is distributed under the MIT license.

Summary of the pyannote/speaker-diarization-3.0 model card, 2026-10-01

Hugging Face ↗Paper ↗