appautomaton/openmoss-ttsd-mlx

OpenMOSS TTSD — MLX

This repository contains an MLX-native int8 conversion of OpenMOSS TTSD for multi-speaker dialogue generation on Apple Silicon.

It is intended for local multi-speaker speech generation with mlx-speech, without a PyTorch runtime at inference time.

Variants

Path	Precision
mlx-int8/	int8 quantized weights

Model Details

• Developed by: AppAutomaton
• Shared by: AppAutomaton on Hugging Face
• Upstream model: OpenMOSS-Team/MOSS-TTSD-v1.0
• Task: multi-speaker text-to-speech
• Runtime: MLX on Apple Silicon

How to Get Started

Command-line generation with mlx-speech:

python scripts/generate_moss_ttsd.py \
  --text "[S1] Watson, I think we should go. [S2] Give me one moment." \
  --output outputs/dialogue.wav

Minimal Python usage:

from mlx_speech.generation import MossTTSDelayModel

model = MossTTSDelayModel.from_path("mlx-int8")

Speaker turns are tagged with [S1] and [S2] in the input text.

Notes

• This repo contains the quantized MLX runtime artifact only.
• The conversion keeps the original TTSD architecture and remaps weights explicitly for MLX inference.
• The current runtime path is designed around speaker-tagged dialogue input and shared codec decoding.

Links

• Source code: mlx-speech
• More examples: AppAutomaton

License

Apache 2.0 — following the upstream license published with OpenMOSS-Team/MOSS-TTSD-v1.0.