Img2Img-Translation-Networks
Img2Img-Translation-Networks copied to clipboard
Exact implementation of referenced paper?
Hey guys,
Great work! Just curious, is this repo following the exact proposed algorithm by "Unsupervised Image-to-Image Translation Networks"?
Based on what I'm seeing from train.py, it doesn't seem to train functions E1, E2, G1, G2 and D1, D2 as proposed by the paper, where E1 maps X1 into the latent space where G2 maps from latent space to X2.
Instead it seems to be mapping between the two groups directly?