| Speech in this dataset was synthesized by MAGO (https://huggingface.co/mago-ai) |
| for full-duplex emotion finetuning. |
|
|
| Speech synthesis |
| Audio was synthesized with Boson AI Higgs Audio v3 TTS 4B |
| (also called Higgs TTS 3), using the Transformers port |
| multimodalart/higgs-audio-v3-tts-4b-transformers. |
| Official weights: https://huggingface.co/bosonai/higgs-tts-3-4b |
|
|
| Derived from Higgs TTS 3, licensed from Boson AI USA, Inc. |
| Copyright Boson AI USA, Inc. All Rights Reserved. |
| https://huggingface.co/bosonai/higgs-tts-3-4b/blob/main/LICENSE |
|
|
| Dialogue text |
| User and agent lines were drafted with Qwen2.5-3B-Instruct |
| (Qwen/Qwen2.5-3B-Instruct, Apache-2.0). |
|
|
| Voice prompts |
| voice_prompts/*.wav are short clips from LibriTTS-R train-clean-100. |
| LibriTTS-R is licensed under CC BY 4.0. |
| https://www.openslr.org/141/ |
| Copyright the LibriTTS-R authors. See voice_prompts/manifest.json |
| for the speaker id and the original utterance filename. |
|
|
| Room tone |
| roomtone/classroom.wav is a 24 kHz mono copy of |
| "roomtone empty classroom 01 short" by klankbeeld, |
| Freesound 212202, CC BY. |
| https://freesound.org/people/klankbeeld/sounds/212202/ |
| Credit: sound from https://freesound.org/people/klankbeeld/ |
| See roomtone/ATTRIBUTION.txt. |
|
|
| This dataset is meant to be read by a PersonaPlex / Moshi finetune |
| (nvidia/personaplex-7b-v1, NVIDIA Open Model License; Moshi by Kyutai, |
| CC BY 4.0). Those model licenses govern the weights, not this data. |
|
|
