Technology

Developing an open-source Audio-to-Animation (A2A) model is a challenging endeavor. We provide miners with reference to open-source pre-trained models DiffuseStyleGesture and EDGE. These models are transformer-based diffusion models for animation generation.

We encourage community members to build, train and upload new A2A models based on these, while also welcoming contributions from other model architectures, training methods and training data to enrich and enhance the outcomes.