Starstruk Logo MarkStarstruk

v6.0

Textualize/textualv6.0Aug 26, 2025by snakers4

AI Summary

This major version introduces a new training algorithm that significantly reduces errors on noisy and multi-domain data, while maintaining previous v5 features and improving overall quality metrics.

Key Highlights

  • New v6 VAD training algorithm
  • 16% reduction in errors on noisy data
  • 11% reduction in errors on multi-domain data
  • Quality comparison with TenVAD added

New Features

  • New training algorithm
  • Error reduction on noisy/multi-domain data

Full Release Notes

## New v6 VAD Released

- Improved [quality](https://github.com/snakers4/silero-vad/wiki/Quality-Metrics#silero-vad-vs-old-silero-vad);
- v5 features and improvements kept;
- 16% less errors on noisy real-life data;
- 11% less errors on multi-domain validation;
- Various community contributions;
- Added quality comparison with TenVAD;
- Changed the training algorithm, ideally it should result in higher robustness;
- Metrics on a new (manually noised) community-provided dataset to be added soon;
- Known persisting issues: music with human voice-like instruments, very high pitched voices (artificial, cartoons, small children);


## What's Changed

* Improve documentation. by @EarningsCall in https://github.com/snakers4/silero-vad/pull/553
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/573
* fx https://github.com/snakers4/silero-vad/issues/576 by @adamnsandle in https://github.com/snakers4/silero-vad/pull/579
* Add cpp source based on libtorch by @NathanJHLee in https://github.com/snakers4/silero-vad/pull/578
* fx negative ths bug by @adamnsandle in https://github.com/snakers4/silero-vad/pull/581
* Add haskell example by @qwbarch in https://github.com/snakers4/silero-vad/pull/591
* Add CITATION.cff file for proper citation by @kiwamizamurai in https://github.com/snakers4/silero-vad/pull/601
* Fix/cpp vad context by @OJRYK in https://github.com/snakers4/silero-vad/pull/605
* Specify time resolution when returning speech coordinates in seconds by @b3by in https://github.com/snakers4/silero-vad/pull/627
* Use second coordinates for audio concatenation in collect_chunks and drop_chunks by @b3by in https://github.com/snakers4/silero-vad/pull/626
* Surface drop_chunks in init by @davidrs in https://github.com/snakers4/silero-vad/pull/656
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/669
* fx by @adamnsandle in https://github.com/snakers4/silero-vad/pull/670
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/671
* fx by @adamnsandle in https://github.com/snakers4/silero-vad/pull/672
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/673
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/674
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/675
* Adamnsandle by @adamnsandle in https://github.com/snakers4/silero-vad/pull/676
* Adding additional params to get_speech_timestamps by @shashank14k in https://github.com/snakers4/silero-vad/pull/664
* get rid of hop_size_ratio by @adamnsandle in https://github.com/snakers4/silero-vad/pull/677

## New Contributors

* @EarningsCall made their first contribution in https://github.com/snakers4/silero-vad/pull/553
* @NathanJHLee made their first contribution in https://github.com/snakers4/silero-vad/pull/578
* @qwbarch made their first contribution in https://github.com/snakers4/silero-vad/pull/591
* @kiwamizamurai made their first contribution in https://github.com/snakers4/silero-vad/pull/601
* @OJRYK made their first contribution in https://github.com/snakers4/silero-vad/pull/605
* @b3by made their first contribution in https://github.com/snakers4/silero-vad/pull/627
* @davidrs made their first contribution in https://github.com/snakers4/silero-vad/pull/656
* @shashank14k made their first contribution in https://github.com/snakers4/silero-vad/pull/664

**Full Changelog**: https://github.com/snakers4/silero-vad/compare/v5.1.2...v6.0