Open SURABHI-GUPTA opened 3 years ago
It will take significant amount of time, for example generation all videos for vox test set is 8 hours. Also training loss has pretty good correlation with final accuracy, so intermediate validation is not really useful.
Sorry, but your code makes no sense to me. You can found how validation is implemented in reconstruction.py
@AliaksandrSiarohin It seems, validation is not done along with the training pipeline. Is there any specific reason for this? What changes do we need to do for adding that to the training pipeline?