Model guides¶

Every registered model has one focused page with the same six sections: overview, quickstart, capabilities, checkpoint and license status, optimization and training, and public API. Hub-backed models also link to a dedicated Colab notebook.

Generated by scripts/generate_model_pages.py. Edit registry and contract metadata, then rerun the generator.

Text to speech¶

Model	Default checkpoint	Training	Notebook
`orpheustts`	`canopylabs/orpheus-3b-0.1-ft`	`native`	Colab
`dia`	`nari-labs/Dia-1.6B-0626`	`native`	Colab
`vui`	`vui-abraham-100m.pt`	`preprocessed`	—
`chatterbox`	`ResembleAI/chatterbox`	`custom`	Colab
`kokoro`	`hexgrad/Kokoro-82M`	`preprocessed`	Colab
`echo`	`jordand/echo-tts-base`	`preprocessed`	Colab
`conversationtts`	`AudioFoundation/SpeechFoundation`	`native`	Colab
`llasa`	`HKUSTAudio/Llasa-1B-Multilingual`	`native`	Colab
`cosyvoice`	`FunAudioLLM/Fun-CosyVoice3-0.5B-2512`	`custom`	Colab
`f5tts`	`F5TTS_v1_Base`	`preprocessed`	—
`gptsovits`	`lj1995/GPT-SoVITS`	`preprocessed`	Colab
`melotts`	`EN`	`preprocessed`	—
`openvoice`	`myshell-ai/OpenVoiceV2`	`custom`	Colab
`outetts`	`OuteAI/Llama-OuteTTS-1.0-1B`	`preprocessed`	Colab
`parlertts`	`parler-tts/parler-tts-mini-v1`	`native`	Colab
`styletts2`	No default; pass a compatible Hub ID or local directory.	`preprocessed`	—
`mosstts`	`OpenMOSS-Team/MOSS-TTS-v1.5`	`native`	Colab
`qwen3tts`	`Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice`	`preprocessed`	Colab
`irodoritts`	`Aratako/Irodori-TTS-500M-v3`	`native`	Colab
`zonos`	`Zyphra/Zonos-v0.1-transformer`	`preprocessed`	Colab
`zonos2`	`Zyphra/ZONOS2`	`native`	Colab
`voxcpm`	`openbmb/VoxCPM2`	`native`	Colab
`omnivoice`	`k2-fsa/OmniVoice`	`native`	Colab
`higgstts`	`bosonai/higgs-tts-2-3b-base`	`native`	Colab
`xtts`	`coqui/XTTS-v2`	`preprocessed`	Colab
`vibevoice`	`microsoft/VibeVoice-Realtime-0.5B`	`preprocessed`	Colab
`fishtts`	`fishaudio/s2-pro`	`preprocessed`	Colab
`csm`	`sesame/csm-1b`	`native`	Colab
`neutts`	`neuphonic/neutts-2e`	`native`	Colab
`supertonic`	`Supertone/supertonic-3`	`preprocessed`	Colab
`inflecttts`	`owensong/Inflect-Micro-v2`	`preprocessed`	Colab
`bark`	`suno/bark-small`	`preprocessed`	Colab
`speecht5`	`microsoft/speecht5_tts`	`native`	Colab
`vits`	`facebook/mms-tts-eng`	`preprocessed`	Colab

Automatic speech recognition¶

Model	Default checkpoint	Training	Notebook
`asr_transformers`	`openai/whisper-small`	`native`	Colab
`asr_whisper`	`openai/whisper-large-v3-turbo`	`native`	Colab
`asr_tiron`	`Trelis/tiron`	`native`	Colab
`asr_qwen3`	`Qwen/Qwen3-ASR-0.6B`	`native`	Colab
`asr_vibevoice`	`microsoft/VibeVoice-ASR-HF`	`native`	Colab
`asr_granite_speech`	`ibm-granite/granite-speech-4.1-2b`	`native`	Colab
`asr_parakeet_tdt`	`nvidia/parakeet-tdt-0.6b-v3`	`native`	Colab
`asr_nemotron`	`nvidia/nemotron-3.5-asr-streaming-0.6b`	`native`	Colab
`asr_cohere`	`CohereLabs/cohere-transcribe-03-2026`	`native`	Colab
`asr_medasr`	`google/medasr`	`native`	Colab
`asr_wav2vec2`	`facebook/wav2vec2-base-960h`	`native`	Colab
`asr_hubert`	`facebook/hubert-large-ls960-ft`	`native`	Colab
`asr_wavlm`	`patrickvonplaten/wavlm-libri-clean-100h-base-plus`	`native`	Colab
`asr_moonshine`	`UsefulSensors/moonshine-tiny`	`native`	Colab
`asr_seamless_m4t_v2`	`facebook/seamless-m4t-v2-large`	`native`	Colab
`asr_faster_whisper`	`openai/whisper-small`	`native`	Colab
`asr_whisperx`	`openai/whisper-small`	`native`	Colab
`asr_openai_whisper`	`openai/whisper-small`	`native`	Colab
`asr_nemo`	`nvidia/nemo/stt_en_quartznet15x5`	`native`	—
`asr_speechbrain`	`speechbrain/asr-crdnn-rnnlm-librispeech`	`native`	Colab
`asr_funasr`	`iic/SenseVoiceSmall`	`native`	Colab
`asr_espnet`	`espnet/shinji-watanabe-librispeech_asr_train_asr_transformer_e18_raw_bpe_sp_valid.acc.best`	`native`	Colab
`asr_wenet`	`wenet/gigaspeech-u2pp-conformer`	`native`	—

Voice activity detection¶

Model	Default checkpoint	Training	Notebook
`vad_transformers`	No default; pass a compatible Hub ID or local directory.	`native`	—
`vad_silero`	`safestack/silero-vad`	`native`	Colab
`vad_webrtc`	`webrtc-vad`	`inference-only`	—
`vad_pyannote`	`pyannote/voice-activity-detection`	`native`	Colab
`vad_speechbrain`	`speechbrain/vad-crdnn-libriparty`	`native`	Colab
`vad_nemo`	`nvidia/Frame_VAD_Multilingual_MarbleNet_v2.0`	`native`	Colab
`vad_funasr`	`funasr/fsmn-vad`	`native`	Colab
`vad_auditok`	`auditok-energy-vad`	`inference-only`	—
`vad_sherpa_onnx`	`safestack/silero-vad`	`native`	Colab
`vad_pyannote_segmentation`	`pyannote/segmentation-3.0`	`native`	Colab
`vad_pyannote_brouhaha`	`pyannote/brouhaha`	`native`	Colab