I trained a model with an accented voice on weights.gg, but the output sounds indistinguishable from generic AI voices found in standard libraries. I’m still exploring the RVC model landscape, but I initially considered Weights.gg as a potential alternative until I could run my own voice modeling using open-source GitHub tools. Despite training my model with the recommended amount of voice data, the results bear little resemblance to the original voice I captured. Does anyone have any tips since I'm new to AI voice generation? I've noticed most pre-made ones are absolutely terrible as well, but I have seen voice generation indistinguishable from the actual characters/celebrities demonstrated before elsewhere so I know the tech is there somewhere, and I want to learn how to use it.
#Weights.gg models aren't good?
nimble orbit