Here are two samples I did with both of them singing together.

RVC v1?

Someone already did a RVC V2 version of Princess Celestia so I was just chose to do RVC V1

You could have done V2 48k.

I don't think that one was done before.

Isn't that option more likely to overtrain sooner?

Besides, I'm just someone who's more likely to switch between both RVC V1 and V2 more often than not

Not really.

If you keep a good eye on the loss/g/total like all rvc models, you should have no problem and it usually sounds better with v2.

But thats just me.

Its up to you honestly if you want to try v2 48k.

Meh I'll try sometime later on another random model But I just don't use RVC V2 that much since it takes forever for import everything each time abd for the Tensorboard and Gradio link to load up

Do you want to sing the zaboomafoo theme song with me