**download link:** https://huggingface.co/lyery/mymodels/resolve/main/wolfver1.zip?download=true

guess who managed to rip the audio files from the game hehe
anyways, the dataset is him talking very deep, monotone and serious so this model can't sing, but works nice in realtime (works better if you talk in japanese and have average japanese male voice) and inferencing speech
he has totally different set of voice lines vs his cbt1 voice lines, but the voice actor seems to be the same (Kobayashi Chikahiro), the new ones are in better quality compared to cbt1 lines (tho still compressed because thats how game files works xd)

**Info:**
Sample rate: 32k
VA: Kobayashi Chikahiro
Pretrain: Original
Precision:Fp32
Pitch extraction: rmvpe
Batch size: 32
Dataset: 2:25 minutes of ripped audio files, i didn't included all of his voicelines because some of them had reverb, and while i know how to remove reverb, raw/barely edited audios are always gonna be better, also i didnt include his grunts because i noticed game characters models sound better without them (the funny huh hah sounds lol)

Additional Details:

Tags: No tags available

Download Link: https://huggingface.co/lyery/mymodels/resolve/main/wolfver1.zip?download=true