Accelerating Flow-Matching-Based Text-to-Speech via Empirically Pruned Step Sampling

Re-assign community