Really great! I've been using diffusion as well to create sample libraries. My angle is to train models strictly on chord progression annotated data as oppose

Really great! I've been using diffusion as well to create sample libraries. My angle is to train models strictly on chord progression annotated data as opposed to the human descriptions so they can be integrated into a DAW plugin. Check it out: https://signalsandsorcery.org/ https://signalsandsorcery.org/