Download Latest Version
v0.5.1 source code.zip (7.4 MB)
Get an email when there's a new version of MLX-Audio
| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| README.md | 2026-08-31 | 2.3 kB | |
| v0.5.1 source code.tar.gz | 2026-08-31 | 6.9 MB | |
| v0.5.1 source code.zip | 2026-08-31 | 7.4 MB | |
| Totals: 3 Items | 14.3 MB | 8 | |
What's Changed
- fix(tests): put ecapa_tdnn models in eval mode before batch=1 forward… by @Lazarus-931 in https://github.com/Blaizzy/mlx-audio/pull/896
- fix(qwen3_tts): stop exempting EOS from top-k/top-p sampling filters by @Lazarus-931 in https://github.com/Blaizzy/mlx-audio/pull/897
- fix(qwen3_tts): reject unsupported voice on Base models instead of silently ignoring it by @Lazarus-931 in https://github.com/Blaizzy/mlx-audio/pull/895
- Implement Breeze-TTS-2 by @SharkyRawr in https://github.com/Blaizzy/mlx-audio/pull/911
- fix(qwen3_tts): window ICL repetition penalty to stop pace acceleration by @Lazarus-931 in https://github.com/Blaizzy/mlx-audio/pull/914
- fix(audio_io): decode CAF, and route mp4/m4b paths through ffmpeg by @Lazarus-931 in https://github.com/Blaizzy/mlx-audio/pull/915
- Add Phonon-1 STT models by @lucasnewman in https://github.com/Blaizzy/mlx-audio/pull/923
- ci: automate PR contributor reminders by @Lazarus-931 in https://github.com/Blaizzy/mlx-audio/pull/924
- Fix: use periodic Hann window in log-mel spectrogram by @monai in https://github.com/Blaizzy/mlx-audio/pull/920
- Fix: honor repetition_penalty in Whisper decoding options by @monai in https://github.com/Blaizzy/mlx-audio/pull/918
- Fix: require low avg_logprob before skipping fallback on no-speech by @monai in https://github.com/Blaizzy/mlx-audio/pull/919
- Add streaming generation for VibeVoice by @lyonsno in https://github.com/Blaizzy/mlx-audio/pull/925
- fix(stt): process every Qwen3-ASR list input by @ainergiz in https://github.com/Blaizzy/mlx-audio/pull/929
- arktts: support Audio8-TTS-Preview-0.1b (Falcon-H1 slow backbone) by @xocialize in https://github.com/Blaizzy/mlx-audio/pull/933
- Add Granite Speech 5.0 STT model by @lucasnewman in https://github.com/Blaizzy/mlx-audio/pull/936
- Bump version to 0.5.1 by @lucasnewman in https://github.com/Blaizzy/mlx-audio/pull/937
New Contributors
- @SharkyRawr made their first contribution in https://github.com/Blaizzy/mlx-audio/pull/911
- @monai made their first contribution in https://github.com/Blaizzy/mlx-audio/pull/920
- @ainergiz made their first contribution in https://github.com/Blaizzy/mlx-audio/pull/929
Full Changelog: https://github.com/Blaizzy/mlx-audio/compare/v0.5.0...v0.5.1
