Skip to content

OmniVoice-bfloat16

mlx-community/OmniVoice-bfloat16
Open on Hugging Face
Parameters
613M
Weights (est.)
≈ 1.2 GB
Mac memory (est.)
8 GB+

Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.

Details

Base model
k2-fsa/OmniVoice (finetune)
Vendor
k2-fsa
Task
Text to speech
Quantisation
Not stated
Licence
apache-2.0
Downloads
1.7K in the last 30 days · 4.3K all time · 2 likes
Published
2026-05-15 · updated 2026-05-30 (5mo ago)
Languages
multilingual
Tags
omnivoice, apple-silicon, voice-cloning

Other MLX builds of k2-fsa/OmniVoice

Build Quantisation Weights Mac memory Downloads
OmniVoice-4bit 4-bit ≈ 192 MB 8 GB+ 348
OmniVoice-8bit 8-bit ≈ 670 MB 8 GB+ 963
OmniVoice-bfloat16 (this one) — ≈ 1.2 GB 8 GB+ 1.7K
OmniVoice — ≈ 2.5 GB 8 GB+ 315
OmniVoice-fp32 fp32 ≈ 2.5 GB 8 GB+ 39