Scaling Rich Style-Prompted Text-to-Speech Datasets

CLIP
VLM
Re-assign community