Singing Voice Synthesis

via “zero-shot voice cloning with minimal reference audio”

text-to-speech model by undefined. 6,61,227 downloads.

Unique: Uses flow matching (continuous normalizing flows) instead of discrete diffusion steps, reducing inference steps from 100+ to 20-30 while maintaining voice fidelity; integrates speaker embeddings via cross-attention rather than concatenation, enabling smoother voice interpolation and style transfer

vs others: Faster inference than XTTS-v2 (2-5s vs 5-10s) with comparable voice quality while requiring less reference audio than Vall-E or YourTTS