Model Details
Voice of Jacob Hopkins as Gumball Watterson in Season 3 of the cartoon The Amazing World of Gumball.
Model Description
<!-- Provide a longer summary of what this model is. -->
- Developed by: ijik-loker
- Model type: Retrieval-based Voice Conversion (RVC)
- Language(s): English
Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
Used in the popular Retrieval-based Voice Conversion WebUI via inference or real-time using Voice Changer. The index file should be used alongside the model.
Training Details
Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
Voice clips dataset total duration
v1 model: 11min 18s
v2 model: 26min 50s
Trained using these episodes from Season 3:
- The Boss
- The Move
- The Burden
- The Bros
- The Countdown
- The Nobody
- The Fraud
- The Void
- The Name
- The Extras (1 line)
- The Oracle
- The Safety
- The Procrastinators
- The Puppy
- The Recipe
- The Society
- The Spoiler
Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
- Remove noise using Ultimate Vocal Remover 5 UVR-DeNoise.
- Extract vocals using RVC Web UI HP5-主旋律人声vocals+其他instrumentals.pth.
- Remove echo and reverb using Ultimate Vocal Remover 5 UVR-DeEcho-DeReverb.
- Manually diarise voices in Audacity using labels.
- Export multiple to .wav by labels.
- Train using RVC
- Target Sample Rate: 48k
- Version: v2
- Total training epochs: 200
- Base model G: f0G48k.pth
- Base model D: f0D48k.pth
Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
Summary
v1 seems to perform just fine. The v2 voice sounds coarse at times.
