Both models are RVCv2 - RVMPE - 48k, both confirmed working individually well on Ilaria but I am unable to merge them, and I do not have tensorflow up at the moment.
60% McCartney
40% Oliver Tree
Goal is to have a model that is 60/40 that I can use in Ilaria RVC, but doesnt sound like two voices harmonizing and isn't robotic or squawky sounding.
I'll put links to the source models in the replies.