BonFrame JoyVASA ONNX runtime pack

ONNX export used by BonFrame's local Apple Silicon lip-redraw pipeline. The pack converts 16 kHz speech audio into JoyVASA motion coefficients; FasterLivePortrait-MLX performs portrait rendering separately.

Derived from jdh-algo/JoyVASA (MIT). The audio feature graph contains the pinned Chinese HuBERT encoder.

Files:

joyvasa_audio_feat.onnx
joyvasa_denoise_step.onnx
joyvasa_meta.json
joyvasa_runtime.json
joyvasa_runtime.npz

Export source: scripts/exportJoyVasaOnnx.py in BonFrame.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support