Fast and High-Quality Auto-Regressive Speech Synthesis via Speculative Decoding

Open in new window