A well-trained 48K Youth Sound RVC model

This training uses a new base mold, which feels much better, especially in laughter and breathing, which are just a bit better than the default template.

If you're interested in this base model, below is the Miaoyin fine-tuning base film.

https://apis.klrvc.com/sybz/2744.html

This time, the youth vocal segment was a fusion training of multiple male vocal tones. This does not refer to model fusion, but rather to fusing similar timbres into training, achieving excellent results and excelling in both high and low notes.

Below is the loss curve chart

Note that since the original input has its own strengths, the output effect will also have its own strengths. There's no guarantee you'll have the same effect as I demonstrated. This requires you to adjust the pitch or hope your tone is close to the inference tone. I specifically included this demonstration before the inference.

So, if you mind this issue, after all, RVC is different from TTS.