PyTorch-SRGAN
PyTorch-SRGAN copied to clipboard
aitorzip
A modern PyTorch implementation of SRGAN
Shouldn't the upsample blocks of the `Generator` in `models.py` be added per powers of 2? The current implementation adds them per multiples of 2, which will only work for upsample=2...
I have a custom dataset of pavement distresses with different types. I would like to know if it is possible to generate or synthesize more training data by utilizing your...
Hi, Thank you for your work. Could you please help me with the problem? ``` Generator pre-training Traceback (most recent call last): File "", line 1, in File "C:\Users\XerCis\AppData\Local\Programs\Python\Python36\lib\multiprocessing\spawn.py", line...
When I tried to use "./test --cuda", it can not work. And it prints " KeyError: 'unexpected key "upsample2.conv.weight" in state_dict'" Can the code work?
Hi, Thx for sharing the code. I'm trying to use the SRGAN with black and white images. The first problem I ran into was that VGG expects 3 channels. The...
Hi, I just changed you code to fit my own dataset. But I got some problem about my training loss... `[4/100][2112/2112] Discriminator_Loss: 0.9368 Generator_Loss (Content/Advers/Total): 0.0021/1.2429/0.0033` `[5/100][2112/2112] Discriminator_Loss: 0.8507 Generator_Loss...
The initial generator pre-training using MSE loss is to initialize the weights properly?
How did you get 96X96 high resolution REAL images for comparison? My understanding was that we take the original dataset (cifar-10, consisting images of dimension 32X32), down-sample the images to...
Hi, i run your code, and the log file store in logs/cifar100-16-0.0001-0.0001/events.out.tfevents.1510472514.nlpr. And then i run the command line: tensorboard --logdir=logs/cifar100-16-0.0001-0.0001, the output result: TensorBoard 0.1.4 at http://nlpr:6006 (Press CTRL+C...
