Re:Generating six-pack abs with Tensorflow
by rayheberer on Wednesday October 18, 2017 @09:43AM
(#15499044)
Great question! My intuition tells me that L1 regularization could indeed be impacting the performance of the model, but for me personally, the only way to find out for certain would be to test this empirically.
Do you think that the L1 regularization used in pix2pix - which is intended to preserve low-frequency spacial patterns in the image pairs - will ultimately end up limiting how well cGANs can learn the body transformation image translation task?