search

sp-uhh / storm

StoRM: A Diffusion-based Stochastic Regeneration Model for Speech Enhancement and Dereverberation
MIT License
164 stars 22 forks source link

48k checkpoint #6

Closed adeelabbas closed 1 year ago

adeelabbas commented 1 year ago

Hi, Do you have a checkpoint that's trained on 48k audio input?

jmlemercier commented 1 year ago

Hi Adeel, unfortunately not: however, you should be able to easily reproduce the conditions using the original VCTK with 48kHz sampling frequency and tweaking the filters lengths / number of channels. We designed experiments for 32kHz audio and it seemed to work okay: cannot share the checkpoints unfortunately as the data has copyrights. Best, Jm