mlx-community/OmniVoice
Text-to-Speech • 0.6B • Updated
by k2-fsa, converted to MLX
Note Default MLX release. Dense fp32 export.
Note 4bit quantized MLX variant.
Note 8bit quantized MLX variant.
Note Dense fp32 reference export.
OmniVoice text-to-speech demo using the MLX backend
Note Interactive Gradio demo for the OmniVoice MLX backend.
Note Dense bfloat16 MLX reference export.