implementation problems here

[Ledzy]

I found two problems about the code.
First, the reconstruction loss should be L1 loss instead of L2, according to cycleGAN's paper.
Second, the generator and discriminator should update simutaneously, while in the code it firstly update discriminator and then use updated discriminator to update generator, which is wrong.