hzzone / du-gan Goto Github PK

你好，你在文章中写到，参考了WGAN-VGG , Q-AE 等几个对比算法的官方实现，但是我没有找到，能分享相关官方代码吗，非常感谢！

why is the test results have two different values like ema_psnr and psnr?
and what is the term ema indicates?

hey, could you please explain to me that how can I represent the training hyper-parameters in terms of epoch because in the paper its mentioned that "During training, we trained the model with a maximum
of 100K iterations and with a mini-batch of size 64 " so that means number of iterations are equal to number of epoch ?
thanks

你好，师兄。我在复现你工作的时候，出现了这个问题。
File "/media/data/sdd_d/yyc/DU-GAN-master/utils/dataset.py", line 12, in init
data = torch.from_numpy(np.load(npy_root).astype(np.float32) - 1024)
错误显示在这里，恳请处理方法！

The network architecture of U-net as the discriminator is related to "U-net: Convolutional networks"
for biomedical image segmentation, "is there a difference in the U-net architecture?

can you tell me how do you prepare the csv files of the datasets and if there is any code for preparing them please share it to me

Thanks for your creative work! May I ask whether the input of the image discriminator and the gradient discriminator is a 2D image or a 3D volume? If I want to try to use the three-dimensional volume as the input of the discriminator, what should I do?

Hello author, regarding the input train_id.csv and test_id.csv of the gen_data file, they are just a set of numbers, and the downloaded data set is just a file after decompression. Is there any problem in it? Looking forward to your reply

is the testing also done parallely using the training shell

Hello, could you help explain how the results of the loss function are saved and how graphs may be extracted from them?