Originally published at: Speeding Up Text-To-Speech Diffusion Models by Distillation | NVIDIA Technical Blog
Every year, as part of their coursework, students from the University of Warsaw, Poland get to work under the supervision of engineers from the NVIDIA Warsaw office on challenging problems in deep learning and accelerated computing. We present the work of three M.Sc. students—Alicja Ziarko, Paweł Pawlik, and Michał Siennicki—who managed to significantly reduce the…
Reducing distillation steps for 5x speed up in latency without compromising speech quality is cool, but will it be enough for real-time Diffusion TTS applications? Can we reach that level? If so, how?