Nowadays, major productivity apps seem to demand a monthly payment. This "subscription creep" is annoying, but there’s a ...

This repository provides a baseline for the CCF-AATC 2025 Challenge Track 1, which aims to do speech restoration under multiple distortions: Acoustic Degradation: (Noise & Reverberation) Signal Chain ...

(1) We released the 50 diffusion steps model (instead of 1000 steps) which runs 20X faster with comparable results. (2) Calling CLIP just once and caching the result runs 2X faster for all models.