From bcdba2d38e9165b8f9b91ce49b3e0059d1ef24da Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" <41898282+github-actions[bot]@users.noreply.github.com> Date: Wed, 25 Mar 2026 21:33:17 +0000 Subject: [PATCH 1/2] chore(wavekat-vad): release v0.1.11 --- Cargo.lock | 2 +- crates/wavekat-vad/CHANGELOG.md | 6 ++++++ crates/wavekat-vad/Cargo.toml | 2 +- 3 files changed, 8 insertions(+), 2 deletions(-) diff --git a/Cargo.lock b/Cargo.lock index f0d21c9..9cdf551 100644 --- a/Cargo.lock +++ b/Cargo.lock @@ -2831,7 +2831,7 @@ dependencies = [ [[package]] name = "wavekat-vad" -version = "0.1.10" +version = "0.1.11" dependencies = [ "criterion", "hound", diff --git a/crates/wavekat-vad/CHANGELOG.md b/crates/wavekat-vad/CHANGELOG.md index 4956017..36e5a65 100644 --- a/crates/wavekat-vad/CHANGELOG.md +++ b/crates/wavekat-vad/CHANGELOG.md @@ -7,6 +7,12 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 ## [Unreleased] +## [0.1.11](https://github.com/wavekat/wavekat-vad/compare/v0.1.10...v0.1.11) - 2026-03-25 + +### Added + +- upgrade silero backend to v6.2.1 ([#42](https://github.com/wavekat/wavekat-vad/pull/42)) + ## [0.1.10](https://github.com/wavekat/wavekat-vad/compare/v0.1.9...v0.1.10) - 2026-03-25 ### Other diff --git a/crates/wavekat-vad/Cargo.toml b/crates/wavekat-vad/Cargo.toml index 4a2ac61..79385f9 100644 --- a/crates/wavekat-vad/Cargo.toml +++ b/crates/wavekat-vad/Cargo.toml @@ -1,6 +1,6 @@ [package] name = "wavekat-vad" -version = "0.1.10" +version = "0.1.11" edition = "2021" description = "Unified voice activity detection with multiple backends" license = "Apache-2.0" From d94e150ce9134604451406833358c164cf1945c8 Mon Sep 17 00:00:00 2001 From: "github-actions[bot]" Date: Wed, 25 Mar 2026 21:34:33 +0000 Subject: [PATCH 2/2] docs: update benchmark table --- README.md | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 8a7e8e2..f08647d 100644 --- a/README.md +++ b/README.md @@ -43,14 +43,14 @@ wavekat-vad = { version = "0.1", features = ["webrtc", "silero", "ten-vad", "fir Performance measured against the [TEN-VAD testset](https://github.com/TEN-framework/ten-vad/tree/main/testset) — 30 audio files from LibriSpeech, GigaSpeech, and DNS Challenge with manual speech/non-speech annotations. Threshold: 0.5. -*v0.1.10* +*v0.1.11* | Backend | Precision | Recall | F1 Score | Frame Size | Avg Inference | RTF | |---------|-----------|--------|----------|------------|---------------|-----| -| WebRTC | 0.821 | 0.983 | 0.895 | 480 (30 ms) | 2.7 µs | 0.0001 | -| Silero | 0.938 | 0.938 | 0.938 | 512 (32 ms) | 121.8 µs | 0.0038 | -| TEN-VAD | 0.942 | 0.915 | 0.928 | 256 (16 ms) | 63.2 µs | 0.0040 | -| FireRedVAD | 0.950 | 0.879 | 0.913 | 160 (10 ms) | 631.6 µs | 0.0632 | +| WebRTC | 0.821 | 0.983 | 0.895 | 480 (30 ms) | 2.9 µs | 0.0001 | +| Silero | 0.938 | 0.938 | 0.938 | 512 (32 ms) | 118.4 µs | 0.0037 | +| TEN-VAD | 0.942 | 0.915 | 0.928 | 256 (16 ms) | 51.4 µs | 0.0032 | +| FireRedVAD | 0.950 | 0.879 | 0.913 | 160 (10 ms) | 647.5 µs | 0.0648 | > Accuracy metrics are deterministic; inference times are approximate and vary by hardware. Measured with `--release` on GitHub Actions `ubuntu-latest` runners. Run locally: `make accuracy` or `make bench`