| --- |
| task_categories: |
| - text-to-speech |
| language: |
| - fr |
| - en |
| pretty_name: Voices TTS-Longeval |
| --- |
| # Voices TTS-Longeval |
|
|
| A set a voices for [kyutai-labs/tts_longeval](https://github.com/kyutai-labs/tts_longeval). |
|
|
| Voice samples are taken from: |
| - `libri/`: the [LibriSpeech ASR Corpus](https://www.openslr.org/12), released under CC BY 4.0 license. |
| - `seed/`: taken from [ButedanceSpeech/seed-tts-eval](https://github.com/BytedanceSpeech/seed-tts-eval), |
| the voices seem to come from [CommonVoice](https://commonvoice.mozilla.org/en), released under CC 0 license. |
| - `ntrex_dialogs/en`: taken from [VCTK](https://datashare.ed.ac.uk/handle/10283/3443), released |
| Creative Commons, Attribution 4.0 International. |
| - `ntrex_dialogs/fr`: taken from the [CML-TTS](https://openslr.org/146/) dataset, released under the |
| Creative Commons License: Attribution 4.0 International. |
| |
| All pre-computed embeddings for the DSM TTS are released under the same license as the original waveform. |
