8-bit quantized MLX conversion of k2-fsa/OmniVoice — a single-stage, non-autoregressive discrete-diffusion text-to-speech model with a Qwen3 backbone and broad multilingual coverage (600+ languages). For Apple Silicon, via the speech-swift runtime.

The diffusion backbone is quantized (it dominates latency — 32 forward passes per clip); the codec stays fp16 since it runs once and is already exact there.

The backbone is memory-bandwidth bound at these sequence lengths, so 8-bit weights cut latency with no audible loss. On an M5 Pro: RTF ≈ 0.18 (≈ 5.5× faster than real-time), versus ≈ 0.25 at fp32, and the ASR roundtrip stays at 0% WER.

Precision tiers:

Loaded by the OmniVoiceTTS module in speech-swift. The quantization block in config.json (group_size: 64, bits: 8) tells the loader to swap the Linear/Embedding layers to their quantized form before loading the weights.

Apache-2.0, following the upstream model.