Reproduce “MeanVoiceFlow2: Joint Optimization of Mean Flow and Content Encoder for Fast One-Step Zero-Shot Voice Conversion” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Takuhiro Kaneko, Hirokazu Kameoka, Kou Tanaka, Yuto Kondo
MeanVoiceFlow2 addresses the speed bottleneck left by a fixed content encoder in one-step zero-shot voice conversion. It jointly trains a content encoder and converter under a MeanVoiceFlow teacher, using reconstruction and teacher-guided augmentation. On VCTK, naturalness rose from 3.76±0.09 to 3.93±0.10, while RTF fell from 0.0072 to 0.00084, with comparable speaker similarity; the comparison covers measured components on one GPU.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.