Ensemble Configuration of TFC-TDF-UNet v3 on MUSDB18-HQ #8
Description
Hello, I am trying to do time measurement on TFC-TDF UNet v3 and I would like to ask some questions regarding the ensemble configuration for the "additional model for MUSDB" mentioned in the paper SOUND DEMIXING CHALLENGE 2023 MUSIC DEMIXING TRACK TECHNICAL REPORT:
TFC-TDF-UNET V3.
Did this model use 3 random seeds and the final output is averaged? And did the results further average with demucs? And I am wondering if it is possible to have the configs/checkpoints of this model? Thanks.