v6.2

RightNow-AI/openfangv6.2Nov 6, 2025by snakers4

AI Summary

This release reworks the VAD training paradigm and improves stability on out-of-distribution data, specifically targeting edge cases involving unusual voices and lower quality audio.

Key Highlights

Reworked VAD training paradigm
Significant quality improvements on edge cases (child, cartoon, muted voices)
Higher stability on OOD and rare data
Type hint fix for min_silence_at_max_speech

New Features

VAD training paradigm rework
Improved stability on rare data
Enhanced handling of muted voices and speech
Fix for phone call quality data

Full Release Notes

- VAD training paradigm reworked;
- Overall slight quality improvement (no metrics update);
- Higher stability on OOD / rare / strange / unique data;
- Significant quality improvements on various known edge cases:
  - Unusual voices
  - Child voices
  - Cartoon voices
  - Muted voices
  - Muted speech
  - Lower quality phone calls

## What's Changed

* Fix type hint for min_silence_at_max_speech (float -> int) by @Purfview in https://github.com/snakers4/silero-vad/pull/714
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/717
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/719

**Full Changelog**: https://github.com/snakers4/silero-vad/compare/v6.1...v6.2