48K Wen Qing Ai Senior, a male RVC model with a decent vocal range

Since some users can't access the 123 drive, I've specially placed it here for you to download and use.

This model was trained on a raw dataset with 48K depth 16 minutes, about 25 minutes long, all very clean, and well-optimized breathing sounds. As for laughter and screaming, although the dataset showed some performance, after several rounds of inference, it still performed well at certain times.

we used AD Senior, Warm Mature Voice RVC model This model serves as inference input (since the two tones differ only in gender), and the two perform quite similarly.

everyone's original input is different, so the output results will vary. You can adjust the tone based on the sound I reasoned to achieve the best effect.

The model algorithm uses the RMVPE algorithm, with 200 training steps, supporting singing. I tested several songs in singing deduction, and the one with the best effect so far is the reasoning one. When covering, I recommend choosing clear singing styles, as ancient-style songs are not very suitable.