A collection of ComfyUI nodes related to audio.
Current nodes:
- Bark
- Load models
- Encode prompt
- Generate
- Semantic
- Coarse
- Fine
- Load/save speaker npz, put speaker .npz files in
models/bark/speakers - HuBert model loader
- HuBert vectorizer (with output layer option, default 9, main voice quantizer was trained for layer 9)
- HuBert quantizer for creating npz from audio https://github.com/gitmylo/bark-voice-cloning-HuBERT-quantizer
- Encodec
- Load models
- Encode audio
- Decode audio
- Coarse-ify codebooks (for creating custom npz files for Bark)