audio-cpp/audio.cpp-gguf

audio.cpp GGUF Model Packages

This directory contains audio.cpp-native GGUF conversions of multiple speech models. These files are intended for use with audio.cpp.

For conversion details, supported layouts, direct-file loading, sidecar embedding, and the latest compatibility notes, see the audio.cpp GGUF guide:

• https://github.com/0xShug0/audio.cpp/blob/main/docs/gguf.md

!!! Automated audio checks are intentionally strict and may flag length, log-mel, or transcript drift that can still sound acceptable to human listeners. Validate the exact file, backend, and route you plan to use. These converted weights are provided as-is; use them at your own risk.

Files

Tested summarizes the current audio.cpp path-test status. See the GGUF guide above for the full matrix and drift notes.

Directory	Files	audio.cpp family	Tested	Original model license
ACE-Step1.5-GGUF	base/ace-step-1.5-base-bf16.gguf, base/ace-step-1.5-base-q8_0.gguf, turbo/ace-step-1.5-turbo-bf16.gguf, turbo/ace-step-1.5-turbo-q8_0.gguf	ace_step	16-bit + Q8 drift	See original model package
Chatterbox-GGUF	chatterbox-f16.gguf, chatterbox-q8_0.gguf	chatterbox	16-bit + Q8 ASR-match drift	MIT
Citrinet-ASR-GGUF	citrinet-asr-q8_0.gguf	citrinet_asr	Q8 pass	CC-BY-4.0
Fish-Audio-S2-Pro-GGUF	fish-audio-s2-pro-bf16.gguf, fish-audio-s2-pro-q8_0.gguf	fish_audio	16-bit + Q8 pass	See original model package
HeartMuLa-GGUF	heartmula-f16.gguf, heartmula-q8_0.gguf	heartmula	16-bit + Q8 drift	Apache-2.0
HTDemucs-GGUF	htdemucs-f16.gguf, htdemucs-q8_0.gguf	htdemucs	16-bit pass, Q8 drift	See original model package
Higgs-Audio-v3-STT-GGUF	higgs-audio-v3-stt-f16.gguf, higgs-audio-v3-stt-q8_0.gguf	higgs_audio_stt	16-bit + Q8 pass	Apache-2.0
Higgs-Audio-v3-TTS-4B-GGUF	higgs-audio-v3-tts-4b-bf16.gguf, higgs-audio-v3-tts-4b-q8_0.gguf	higgs_audio_tts	16-bit + Q8 pass	See original model package
IndexTTS2-GGUF	index-tts2-orig.gguf, index-tts2-f16.gguf, index-tts2-q8_0.gguf	index_tts2	orig + 16-bit pass/drift, Q8 ASR-match drift	bilibili Model Use License Agreement
Irodori-TTS-500M-v3-GGUF	irodori-tts-500m-v3-f16.gguf, irodori-tts-500m-v3-q8_0.gguf	irodori_tts	16-bit pass, Q8 drift	MIT
Irodori-TTS-600M-v3-VoiceDesign-GGUF	irodori-tts-600m-v3-voicedesign-f16.gguf, irodori-tts-600m-v3-voicedesign-q8_0.gguf	irodori_tts	16-bit pass, Q8 drift	MIT
MOSS-TTS-Local-v1.5-GGUF	moss-tts-local-v1.5-bf16.gguf, moss-tts-local-v1.5-q8_0.gguf	moss_tts_local	16-bit pass, Q8 ASR-match drift	Apache-2.0
MOSS-TTS-Nano-100M-GGUF	moss-tts-nano-100m-bf16.gguf, moss-tts-nano-100m-q8_0.gguf	moss_tts_nano	16-bit pass, Q8 ASR-match drift	Apache-2.0
Mel-Band-RoFormer-GGUF	mel-band-roformer-f16.gguf, mel-band-roformer-q8_0.gguf	mel_band_roformer	16-bit + Q8 drift	MIT
MioCodec-25Hz-44.1kHz-v2-GGUF	miocodec-25hz-44khz-v2-orig.gguf, miocodec-25hz-44khz-v2-f16.gguf, miocodec-25hz-44khz-v2-q8_0.gguf	miocodec	orig pass, 16-bit + Q8 drift	MIT
MioTTS-1.7B-GGUF	miotts-1.7b-orig.gguf, miotts-1.7b-bf16.gguf, miotts-1.7b-q8_0.gguf	miotts	orig pass, 16-bit drift, Q8 ASR-match drift	Apache-2.0
Nemotron-3.5-ASR-Streaming-0.6B-GGUF	nemotron-3.5-asr-streaming-0.6b-f16.gguf, nemotron-3.5-asr-streaming-0.6b-q8_0.gguf	nemotron_asr	16-bit pass, Q8 minor filler drift	OpenMDW-1.1
OmniVoice-GGUF	omnivoice-bf16.gguf, omnivoice-f16.gguf, omnivoice-q8_0.gguf	omnivoice	16-bit + Q8 drift	Apache-2.0
PocketTTS-GGUF	english/, german/, italian/, portuguese/, spanish/ each contain bf16 and q8_0 GGUFs	pocket_tts	16-bit pass, Q8 drift	See original model package
Qwen3-ASR-0.6B-GGUF	qwen3-asr-0.6b-f16.gguf, qwen3-asr-0.6b-q8_0.gguf	qwen3_asr	16-bit + Q8 pass	Apache-2.0
Qwen3-ASR-1.7B-GGUF	qwen3-asr-1.7b-f16.gguf, qwen3-asr-1.7b-q8_0.gguf	qwen3_asr	16-bit + Q8 pass	Apache-2.0
Qwen3-ForcedAligner-0.6B-GGUF	qwen3-forced-aligner-0.6b-f16.gguf, qwen3-forced-aligner-0.6b-q8_0.gguf	qwen3_forced_aligner	16-bit + Q8 pass	Apache-2.0
Qwen3-TTS-12Hz-1.7B-Base-GGUF	qwen3-tts-12hz-1.7b-base-orig.gguf, qwen3-tts-12hz-1.7b-base-bf16.gguf, qwen3-tts-12hz-1.7b-base-q8_0.gguf	qwen3_tts	orig pass, 16-bit + Q8 ASR-match drift	Apache-2.0
Qwen3-TTS-12Hz-1.7B-CustomVoice-GGUF	qwen3-tts-12hz-1.7b-customvoice-bf16.gguf, qwen3-tts-12hz-1.7b-customvoice-q8_0.gguf	qwen3_tts	16-bit + Q8 ASR-match drift	Apache-2.0
Qwen3-TTS-12Hz-1.7B-VoiceDesign-GGUF	qwen3-tts-12hz-1.7b-voicedesign-bf16.gguf, qwen3-tts-12hz-1.7b-voicedesign-q8_0.gguf	qwen3_tts	16-bit + Q8 ASR-match drift	Apache-2.0
SeedVC-MLX-GGUF	seed-vc-mlx-orig.gguf, seed-vc-mlx-f16.gguf, seed-vc-mlx-q8_0.gguf	seed_vc	16-bit + Q8 drift	GPL-3.0
Sortformer-Diar-4spk-v1-GGUF	sortformer-diar-4spk-v1-f16.gguf, sortformer-diar-4spk-v1-q8_0.gguf	sortformer_diar	16-bit + Q8 pass	CC-BY-NC-4.0
Stable-Audio-3-Medium-GGUF	stable-audio-3-medium-f16.gguf, stable-audio-3-medium-q8_0.gguf	stable_audio	16-bit + Q8 drift	Stability AI Community License
Stable-Audio-3-Small-Music-GGUF	stable-audio-3-small-music-f16.gguf, stable-audio-3-small-music-q8_0.gguf	stable_audio	16-bit + Q8 drift	Stability AI Community License
Stable-Audio-3-Small-SFX-GGUF	stable-audio-3-small-sfx-f16.gguf, stable-audio-3-small-sfx-q8_0.gguf	stable_audio	16-bit + Q8 drift	Stability AI Community License
Supertonic-3-GGUF	supertonic-3-orig.gguf, supertonic-3-f16.gguf, supertonic-3-q8_0.gguf	supertonic	F32/orig pass; f16 not tested; Q8 unsupported dtype	BigScience Open RAIL-M
Vevo2-GGUF	vevo2-orig.gguf, vevo2-f16.gguf, vevo2-q8_0.gguf	vevo2	orig + 16-bit pass/drift; Q8 mixed route drift	See original model package
VibeVoice-1.5B-GGUF	vibevoice-1.5b-bf16.gguf, vibevoice-1.5b-q8_0.gguf	vibevoice	16-bit pass, Q8 drift	MIT
VibeVoice-ASR-GGUF	vibevoice-asr-f16.gguf, vibevoice-asr-q8_0.gguf	vibevoice_asr	16-bit + Q8 pass	MIT
VoxCPM2-GGUF	voxcpm2-orig.gguf, voxcpm2-bf16.gguf, voxcpm2-q8_0.gguf	voxcpm2	orig pass, 16-bit + Q8 ASR-match drift	Apache-2.0
Voxtral-Mini-4B-Realtime-2602-GGUF	voxtral-mini-4b-realtime-2602-bf16.gguf, voxtral-mini-4b-realtime-2602-q8_0.gguf	voxtral_realtime	16-bit + Q8 pass	Apache-2.0

Q8 Notes

• Chatterbox Q8 is intentionally mixed type. Graph-sensitive scalar, norm, bias, and side tensors stay in non-Q8 types while matmul-compatible weights are quantized.
• PocketTTS Q8 keeps the four flow_lm.flow_net.time_embed.*.mlp.{0,2}.weight tensors in Q8 in addition to the default converter selection. conditioner.embed, cond_embed, and Mimi conv tensors are not forced to Q8 because tested outputs drifted or the current conv path casts quantized conv weights back to F32.

Usage

Pass a GGUF file directly as --model:

audiocpp_cli --task tts --family supertonic --model Supertonic-3-GGUF/supertonic-3-orig.gguf --backend cuda --language en --text "Hello." --voice-id M1 --out out.wav

For ASR:

audiocpp_cli --task asr --family qwen3_asr --model Qwen3-ASR-0.6B-GGUF/qwen3-asr-0.6b-f16.gguf --backend cuda --audio speech.wav --text "" --text-out transcript.txt

License

Each GGUF file is a converted form of its original model. Use and redistribution are governed by the corresponding original model license listed above. Please review the original model card and license terms before using or redistributing any converted weights.