Not working on other datasets #457
Description
The codes for deep domain adaptation seem to not be improving the accuracy when tested on other datasets than provided in the description, the accuracy seems not to improve or even decrease with every epoch, did any one figured out how to fix this?