- facades
- edges2shoes
- edges2handbags
- apple2orange
- zebra2horse
pytorch implementaion(iteration = 8950,batch_size 200) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
pytorch implementaion(iteration = 173850,batch_size= 200) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
pytorch implementaion(iteration = 15450,batch_size= 200) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
Using Pix2pix's unet and combination of LSGAN and L1 For Lconst loss
Can train some dataset for 256 x 256 image size(discogan_pytorch_gpu_torloader_deep.py)
pytorch implementaion(iteration = 23916,batch_size= 25) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
pytorch implementaion(epoch = 106,batch_size= 25) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
pytorch implementaion(epoch = 200,batch_size= 25) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
pytorch implementaion(epoch = 299,batch_size= 25) |
A |
AB |
ABA |
|
|
|
B |
BA |
BAB |
|
|
|
My codes are heavily rely on https://github.com/carpedm20/DiscoGAN-pytorch
Special thanks to carpedm20
- Isola, Phillip, et al. "Image-to-image translation with conditional adversarial networks."
(Full paper: https://arxiv.org/pdf/1611.07004.pdf)
Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks
(arXiv:1703.10593 [cs.CV])
- Learning to Discover Cross-Domain Relations with Generative Adversarial Networks
(arXiv:1703.05192 [cs.CV])
- DualGAN: Unsupervised Dual Learning for Image-to-Image Translation
(arXiv:1704.02510 [cs.CV])
- https://github.com/SKTBrain/DiscoGAN
- https://github.com/carpedm20/DiscoGAN-pytorch
- https://github.com/duxingren14/DualGAN
- https://github.com/junyanz/CycleGAN
- https://github.com/znxlwm/pytorch-CycleGAN
- https://github.com/togheppi/DualGAN