juanmc2005/diart
A python package to build AI-powered real-time audio applications
Leverages speaker segmentation and embedding models with incremental clustering for real-time speaker diarization that improves accuracy as conversations progress. Offers modular pipelines for voice activity detection and transcription, integrates pre-trained models from Hugging Face and Pyannote, and supports custom model integration via ONNX and PyTorch. Provides WebSocket support for web deployment and includes CLI tools for streaming from microphones or audio files.
1,944 stars. No commits in the last 6 months. Available on PyPI.
Stars
1,944
Forks
159
Language
Python
License
MIT
Category
Last pushed
Feb 12, 2025
Commits (30d)
0
Dependencies
20
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/juanmc2005/diart"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related frameworks
felixbur/nkululeko
Machine learning speaker characteristics
claritychallenge/clarity
Clarity Challenge toolkit - software for building Clarity Challenge systems
astorfi/3D-convolutional-speaker-recognition
:speaker: Deep Learning & 3D Convolutional Neural Networks for Speaker Verification
wq2012/awesome-diarization
A curated list of awesome Speaker Diarization papers, libraries, datasets, and other resources.
hitachi-speech/EEND
End-to-End Neural Diarization