-
On the Effectiveness of Least Squares Generative Adversarial Networks
https://arxiv.org/pdf/1712.06391.pdf
-
I want to ask that what the "source_file" in the age_lsgan_transfer.py is. According to the code,it consists of img_name and label
"""
:param filename:
each line in filename is img_name \space labe…
-
I try to implement msg way to cycleGAN for two weeks.
The generator : Unet-withSkipConnection and downsample(k=4, s=2 conv) upsample(Deconv) 8x.
But the D_loss, both D_A and D_B always decreased n…
-
Hi, Prof. Qi,
I am wondering why there is a ReLU layer at the top of discriminator (shown at line 120 in lsgan.lua)? With this ReLU layer, I found the code cannot train the model at all.
thanks,
-
I ran `test.py` with the pre-trained model `0_conv5_lsgan_transfer_g75_0.5f-4_a30`.
For each input images, there are four generated images _1 through _4.
What are the supposed age differences of…
-
Hi there, thanks for putting this repo together! I'm wondering what kind of throughput people are seeing for training? Im getting about 1 iteration every 3 seconds with a batch size of 1. Seems a bit …
-
hello, I noticed that you add WGAN-GP loss in CycleGAN.
I am wondering that if the generator will oscillating during training using wgan loss or wgan-gp loss instead of lsgan loss because the wgan …
-
Hi~
I was wondering why the discriminator doesn't have activation function? As shown in the following code, the real and fake labels are set to 1 and 0 respectively in lsgan mode, what if the outp…
-
Hello,
Thank you for the implementation of the Wasserstein GANs mode and GP loss!
I followed the way proposed here: https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/issues/439
It works fo…
-
## ❓ Questions
In the [paper](https://arxiv.org/abs/2210.13438) 3.4 "Discriminative Loss" section, adversarial loss is constructed as $l_g(\hat{x})=\mathbb{E}[max(0,1-D_k(\hat{x}))]$, but in the [o…