Search results

Qwen3-ASR ignores --max-parallel-segments; decodes chunks sequentially

Status: Open.

#782 In Blaizzy/mlx-audio;

· hhh2210 opened on Jun 7, 2026

feat(stt): SRT/VTT cues are one-per-chunk and too long for subtitles

Status: Open.

#780 In Blaizzy/mlx-audio;

· hhh2210 opened on Jun 6, 2026

Support for Granite 4.1 - Standard and Plus variants?

Status: Open.

#779 In Blaizzy/mlx-audio;

· victorhooi opened on Jun 6, 2026

Feature: Add SenseVoice STT backend — non-autoregressive, 50+ languages, emotion detection

Status: Open.

#753 In Blaizzy/mlx-audio;

· LauraGPT opened on May 30, 2026

Fish S2 giving hallucinated result

Status: Open.

#747 In Blaizzy/mlx-audio;

· oystermove6r opened on May 27, 2026

VoxCPM2-8bit still not loadable in 0.4.3; add voxcpm2 loader support

Status: Open.

#736 In Blaizzy/mlx-audio;

· DrMaks22 opened on May 21, 2026

Stochastic model.generate() hangs in multi-process deployment (cross-process Metal contention, not addressed by PR #668)

Status: Open.

#733 In Blaizzy/mlx-audio;

· maria-oraforge opened on May 20, 2026

UI files not included in package properly

Status: Open.

#723 In Blaizzy/mlx-audio;

· clayrosenthal opened on May 15, 2026

Qwen3TTSBatchSession.step does not stream incrementally or emit final chunk on 0.4.3

Status: Open.

#720 In Blaizzy/mlx-audio;

· terrazorc opened on May 13, 2026

Support MiMo-V2.5-ASR

Status: Open.

#718 In Blaizzy/mlx-audio;

· ailuntx opened on May 12, 2026

OmniVoice generates empty WAVs when audio tokenizer load fails (which is always)

Status: Open.

#714 In Blaizzy/mlx-audio;

· DD-65 opened on May 6, 2026

Voxtral-4B-TTS: expose speaker encoder / add_voice_from_audio for cloning from arbitrary reference WAVs

Status: Open.

#694 In Blaizzy/mlx-audio;

· adrianwedd opened on Apr 30, 2026