Dear Nvidia Jetson community,
We have developed a photorealistic avatar using NVIDIA Jetson AGX Orin, which was used to run Large Language model inference and send audio data to Nvidia RTX-based machine running NVIDIA Omniverse Audio2Face and Unreal Engine software tools. The architecture of our photorealistic avatar is shown below. The code for the project will be provided later. Kudos to Dustin (@dusty_nv) for his work on the llamaspeak project, which was important in making this achievement possible.
The demo video:
Best regards,
Shakhizat