From the ACIDS for at IRCAM - Official implementation of RAVE: A variational autoencoder for fast and high-quality neural audio synthesis (article link) by Antoine Caillon and Philippe Esling.<p>Video of use with ~nn puredata & max/msp object: <a href="https://www.youtube.com/watch?v=dMZs04TzxUI" rel="nofollow">https://www.youtube.com/watch?v=dMZs04TzxUI</a><p>Video showing use of realtime embedded RAVE models on different platforms (Raspberry Pi 4 / Nvidia Jetson Nano): <a href="https://www.youtube.com/watch?v=jAIRf4nGgYI" rel="nofollow">https://www.youtube.com/watch?v=jAIRf4nGgYI</a>