Modelpublic

plaincompute/audio8-tts-preview-0.6b-mx

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face

1likes81downloads

Model Card

Audio8-TTS-Preview-0.6b-bf16 (MLX)

MLX conversion of Audio8/Audio8-TTS-Preview-0.6b (revision 1b17c91) for Apple Silicon — a 0.6B-parameter multilingual text-to-speech model with zero-shot voice cloning and a bundled 44.1 kHz neural codec. And then quantized to Q6 or Q4 for variable layers. Then finally merge codec into the model to get a less than 1GB model file.

—Language model: bf16 (as published upstream)
—Codec: fp32, converted from codec.pth to safetensors with weight-norm folded and conv weights in MLX (channels-last) layout
—Architecture: arktts — a DualAR transformer (24-layer slow AR predicting one semantic token per frame, 4-layer fast AR predicting 10 codec codebooks per frame), inspired by Fish Audio S2 Pro

Conversion parity vs the PyTorch reference (fp32, CPU): unit/block outputs within 1e-4, reference-audio codec encoding 100% code-exact, greedy generation 100% token-exact over the validation utterance, decoded waveform max-abs 7.5e-6.

Usage

Currently can only be used in https://github.com/jason-ni/tf-assistant-app

Supported languages

Cantonese, Chinese, Dutch, English, French, German, Italian, Japanese, Korean, Polish, Spanish (per the upstream preview release).

License

Apache-2.0, following the upstream model. The bundled codec weights are part of the upstream repository and carry the same license.