How much VRAM does OmniVoice need?

Q4_K_M is exactly 659,959,328 bytes (0.61 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

k2-fsa · audio tts

OmniVoice

k2-fsa/OmniVoice

OmniVoice at Q4_K_M is exactly 659,959,328 bytes (0.61 GiB / 0.66 GB) — an effective 8.619 bits per weight, not the nominal 4.

From the file· summed from 2 file(s)From the file· KV per layer

Parameters

613M

Architecture

omnivoice-lm

null layers

Context

—

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q4_K	0.56 GiB	597,547,584	7.804	—	cstr
Q4_K_M2 shards	0.61 GiB	659,959,328	8.619	—	Serveurperso
Q8_02 shards	0.88 GiB	945,284,608	12.345	—	Serveurperso
Q8_02 shards	1.06 GiB	1,137,987,648	14.862	—	cstr
BF162 shards	1.49 GiB	1,603,932,928	20.947	—	Serveurperso
F162 shards	1.52 GiB	1,633,808,896	21.337	—	cstr
F322 shards	2.97 GiB	3,189,948,864	—	—	Serveurperso

Compare with

same modality, comparable size

vs Ace-Step1.5 vs Qwen3-TTS-12Hz-0.6B-Base vs orpheus-3b-0.1-ft vs VieNeu-TTS-0.3B vs neutts-air vs Fun-CosyVoice3-0.5B-2512 All alternatives →Will it fit on my card? →

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 0.32 GiB. The real file is 0.61 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json

Layers

—

Attention heads

—

KV heads

—

Head dim

—

Hidden size

—

Vocab

—

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does OmniVoice need?: Q4_K_M is exactly 659,959,328 bytes (0.61 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of OmniVoice should I use?: Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.