Modelpublic

endo5501/audio.cpp

sourceHugging Facemitupdated 2mo agoView on Hugging Face

0likes55downloads

Model Card

Irodori-TTS model assets for audio.cpp (NovelViewer)

Runtime model assets for the endo5501/audio.cpp fork (Irodori-TTS engine used by NovelViewer). This repository repackages the minimal file set required by the audio.cpp irodori_tts safetensors loader, laid out as sibling directories:

Irodori-TTS-600M-v3-VoiceDesign/
  model.safetensors
  model_config.json
llm-jp-3-150m/
  tokenizer.json
Semantic-DACVAE-Japanese-32dim/
  weights.safetensors   (converted from upstream weights.pth)

Sources and licenses

Asset	Upstream	License
Irodori-TTS-600M-v3-VoiceDesign	Aratako/Irodori-TTS-600M-v3-VoiceDesign	MIT (+ ethical restrictions, see below)
llm-jp-3-150m tokenizer	llm-jp/llm-jp-3-150m	Apache-2.0
Semantic-DACVAE-Japanese-32dim	Aratako/Semantic-DACVAE-Japanese-32dim	MIT

Semantic-DACVAE-Japanese-32dim/weights.safetensors is a format conversion (PyTorch weights.pth → safetensors) of the upstream checkpoint; weights are unmodified.

Ethical restrictions (inherited from Irodori-TTS)

In addition to the MIT license terms, the upstream Irodori-TTS model states ethical restrictions on use (e.g., prohibiting impersonation without consent and unlawful use). See the upstream model card for the authoritative text: https://huggingface.co/Aratako/Irodori-TTS-600M-v3-VoiceDesign