Aller au contenu

Model list

Choose a model to open its dedicated usage, source, training, checkpoint, and public API documentation.

Text to speech

Model Languages Hugging Face ID Training Notebook
Bark de, en, es, fr, … complete audited list on the model page suno/bark-small preprocessed Colab
Chatterbox ar, da, de, el, … complete audited list on the model page ResembleAI/chatterbox custom Colab
ConversationTTS en, zh, yue AudioFoundation/SpeechFoundation native Colab
CosyVoice zh, en, ja, ko, … complete audited list on the model page FunAudioLLM/Fun-CosyVoice3-0.5B-2512 custom Colab
CSM en sesame/csm-1b native Colab
Dia en nari-labs/Dia-1.6B-0626 native Colab
EchoTTS en jordand/echo-tts-base preprocessed Colab
F5TTS en, zh SWivid/F5-TTS preprocessed —
FishTTS zh, en, ja, ko, … complete audited list on the model page fishaudio/s2-pro preprocessed Colab
GPTSoVITS zh, en, ja, ko, yue lj1995/GPT-SoVITS preprocessed Colab
HiggsTTS en, zh, de, ko bosonai/higgs-tts-2-3b-base native Colab
InflectTTS en-US owensong/Inflect-Micro-v2 preprocessed Colab
IrodoriTTS ja Aratako/Irodori-TTS-500M-v3 native Colab
Kokoro en-US, en-GB, es, fr, … complete audited list on the model page hexgrad/Kokoro-82M preprocessed Colab
Llasa zh, en, de, fr, … complete audited list on the model page HKUSTAudio/Llasa-1B-Multilingual native Colab
MeloTTS en, fr, ja, es, zh, ko myshell-ai/MeloTTS-English preprocessed —
MossTTS zh, yue, en, ar, … complete audited list on the model page OpenMOSS-Team/MOSS-TTS-v1.5 native Colab
NeuTTS en neuphonic/neutts-2e native Colab
OmniVoice aae, aal, aao, ab, … complete audited list on the model page k2-fsa/OmniVoice native Colab
OpenVoice en, es, fr, zh, ja, ko myshell-ai/OpenVoiceV2 custom Colab
OrpheusTTS en canopylabs/orpheus-3b-0.1-ft native Colab
OuteTTS en, ar, zh, nl, … complete audited list on the model page OuteAI/Llama-OuteTTS-1.0-1B preprocessed Colab
ParlerTTS en parler-tts/parler-tts-mini-v1 native Colab
Qwen3TTS zh, en, ja, ko, … complete audited list on the model page Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice preprocessed Colab
SpeechT5 en microsoft/speecht5_tts native Colab
StyleTTS2 en-US yl4579/StyleTTS2-LibriTTS preprocessed —
Supertonic en, ko, ja, ar, … complete audited list on the model page Supertone/supertonic-3 preprocessed Colab
VibeVoice en microsoft/VibeVoice-Realtime-0.5B preprocessed Colab
Vits en facebook/mms-tts-eng preprocessed Colab
VoxCPM zh, en, ar, my, … complete audited list on the model page openbmb/VoxCPM2 native Colab
Vui en fluxions/vui preprocessed —
XTTS en, es, fr, de, … complete audited list on the model page coqui/XTTS-v2 preprocessed Colab
Zonos en, ja, zh, fr, de Zyphra/Zonos-v0.1-transformer preprocessed Colab
Zonos2 en, zh, ja, ko, … complete audited list on the model page Zyphra/ZONOS2 native Colab

Automatic speech recognition

Model Languages Hugging Face ID Training Notebook
Cohere ar, de, el, en, … complete audited list on the model page CohereLabs/cohere-transcribe-03-2026 native Colab
ESPnetASR en espnet/shinji-watanabe-librispeech_asr_train_asr_transformer_e18_raw_bpe_sp_valid.acc.best native Colab
FasterWhisper en, zh, de, es, … complete audited list on the model page openai/whisper-small native Colab
FunASR zh, en, ja, ko, yue FunAudioLLM/SenseVoiceSmall native Colab
GraniteSpeech en, fr, de, es, pt, ja ibm-granite/granite-speech-4.1-2b native Colab
Hubert en facebook/hubert-large-ls960-ft native Colab
MedASR en google/medasr native Colab
Moonshine en UsefulSensors/moonshine-tiny native Colab
NeMoASR en Not published / not applicable native —
Nemotron en-US, en-GB, es-US, es-ES, … complete audited list on the model page nvidia/nemotron-3.5-asr-streaming-0.6b native Colab
OpenAIWhisper en, zh, de, es, … complete audited list on the model page openai/whisper-small native Colab
ParakeetTDT en, es, fr, de, … complete audited list on the model page nvidia/parakeet-tdt-0.6b-v3 native Colab
Qwen3ASR ar, yue, zh, cs, … complete audited list on the model page Qwen/Qwen3-ASR-0.6B native Colab
SeamlessM4Tv2 afr, amh, arb, ary, … complete audited list on the model page facebook/seamless-m4t-v2-large native Colab
SpeechBrainASR en speechbrain/asr-crdnn-rnnlm-librispeech native Colab
Tiron en, zh Trelis/tiron native Colab
TransformersASR en, zh, de, es, … complete audited list on the model page openai/whisper-small native Colab
VibeVoice en, zh, es, pt, … complete audited list on the model page microsoft/VibeVoice-ASR-HF native Colab
Wav2Vec2 en facebook/wav2vec2-base-960h native Colab
WavLM en patrickvonplaten/wavlm-libri-clean-100h-base-plus native Colab
WeNetASR en Not published / not applicable native —
Whisper en, zh, de, es, … complete audited list on the model page openai/whisper-large-v3-turbo native Colab
WhisperX en, zh, de, es, … complete audited list on the model page openai/whisper-small native Colab

Voice activity detection

Model Languages Hugging Face ID Training Notebook
AuditokVAD Not text-language conditioned Not published / not applicable inference-only —
FunASRVAD Not text-language conditioned funasr/fsmn-vad native Colab
NeMoVAD Not text-language conditioned nvidia/Frame_VAD_Multilingual_MarbleNet_v2.0 native Colab
PyannoteBrouhahaVAD Not text-language conditioned pyannote/brouhaha native Colab
PyannoteSegmentationVAD Not text-language conditioned pyannote/segmentation-3.0 native Colab
PyannoteVAD Not text-language conditioned pyannote/voice-activity-detection native Colab
SherpaONNXVAD Not text-language conditioned safestack/silero-vad native Colab
SileroVAD Not text-language conditioned safestack/silero-vad native Colab
SpeechBrainVAD Not text-language conditioned speechbrain/vad-crdnn-libriparty native Colab
TransformersVAD Not text-language conditioned Not published / not applicable native —
WebRTCVAD Not text-language conditioned Not published / not applicable inference-only —

Registry access

Registry discovery stays lazy and imports no model runtime.

from voicehub import list_model_specs

for model in list_model_specs():
    print(model.task.value, model.display_name, model.model_type)

Use the training matrix and optimization catalog for compact comparisons.

Generated by scripts/generate_model_pages.py from lazy registry metadata.