Published May 31, 2023 | Version v1

Model weights and pre-processed audio files for Diffiner

  • 1. Sony Group Corporation
  • 2. Sony Research

Description

This repository contains the pre-trained model weights for Diffiner proposed in the paper Diffiner: A Versatile Diffusion-based Generative Refiner for Speech Enhancement by Sony. Furthermore, to run Diffiner same as our paper, the preceding results processed by DCUnet are provided here.

More information about Diffiner including our codes is available in https://github.com/sony/diffiner.

Files

Files (710.4 MB)

Name Size
md5:04b80728ef9a704d2f56114c8ad3628d
657.5 MB Download
md5:180177a502010bb7f312e3cb89bd479f
53.0 MB Download