**download link**: https://huggingface.co/lyery/mymodels/resolve/main/dragoncrepe2.zip?download=true

every League character has their voice lines compressed so much, mp3 compression basically but in .wem format, og crepe (not mangio-crepe) seem to handle mp3 compression a bit better than rmvpe, at least in this dataset og crepe made audible compression lower volume

**model info**
batch size: 4
pitch extraction method: default crepe
hop length: 320
dataset: 6:18 minutes, deleted some samples to avoid potential timbre issues
sample rate: 32k
steps: 2822, fm epoch
va: Georgina Cordova

**tips**
inference crepe with this model, not mangio
in case crepe fails to extract the pitch correctly (aka the voice becoming super distorted), use rmvpe
in case you want to inference mangio-crepe, use a hop length between 128 to 64, mangio inference has more voice cracks than both og crepe and rmvpe so keep in mind that

Additional Details:

Tags: No tags available

Download Link: https://huggingface.co/lyery/mymodels/resolve/main/dragoncrepe2.zip%3Fdownload%3Dtrue?download=true Hosted on Hugging Face

Convert this model in EasyAIVoice

Original Link: https://huggingface.co/lyery/mymodels/resolve/main/dragoncrepe2.zip?download=true (404 not found)

Auto-resolved Link: https://huggingface.co/lyery/mymodels/resolve/main/dragoncrepe2.zip%3Fdownload%3Dtrue?download=true