Note Unofficial demo for E2/F5-TTS, which supports zero-shot voice cloning

Note Spaces of the Week! Note: I did not create the model, just the demo.

Note Unofficial demo for the Moonshine ASR model, an efficient & fast ASR model by Useful Sensors Moonshine ASR: https://github.com/usefulsensors/moonshine

Note Demo for MeloTTS: Multilingual, multispeaker text-to-speech licensed under the MIT license

Note My newest project, a demo of RWKV 4 Music (the MIDI model).

Note My most successful project: an online demo for StyleTTS 2. Reached HF Spaces of the Week and was the most popular Space of the Week. Note: I did not create StyleTTS 2, just the demo.

Note Filtered version of Google's FLEURS dataset (English only), removing ~1/3 of the samples based on an automated quality score.

Note A multilingual dataset of text-phoneme pairs supporting 15 languages. This dataset was created in collaboration with other amazing open source contributors!