super-resolution
super-resolution copied to clipboard
collection of super-resolution models & algorithms
I changed the ‘load_img' function in file ‘dataset.py’ to load RGB images. And the parameter 'num_channels' in 'build_model' function in 'solver.py' to 3. And the 'super_resolve.py' to load the model...
I think the resize input_transform should be bicubic since the default value is bilinear, and centercrop should be changed to RandomCrop() and some of the calculation of the PSNR in...
Please take a look at https://github.com/ChaofWang/Awesome-Super-Resolution and https://github.com/YapengTian/Single-Image-Super-Resolution for other algorithms that you can implement in order to be complete. Also other parts of interest: - https://github.com/xiaomi-automl/FALSR - https://github.com/huangzehao/Super-Resolution.Benckmark -...
Hello, when I run fsrcnn and EDSR, the model does not converge. What's the matter?
the model of EDSR not correct! `Net( (input_conv): Conv2d(1, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (residual_layers): Sequential( (0): ResnetBlock( (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1)) (conv2):...
i found the nEpoch you set is only 20. maybe it is not a suitable setting for training. could you please help me to set a appropriate hyperparas such as...
Which is the best algorithm?
~/super-resolution$ python3 super_resolve.py --model NTIRE2018_x8.pth --input bo-ky-soup.jpg Namespace(input='bo-ky-soup.jpg', model='NTIRE2018_x8.pth', output='test.jpg') Traceback (most recent call last): File "super_resolve.py", line 34, in model = model.to(device) AttributeError: 'collections.OrderedDict' object has no attribute 'to'...
didn't use vgg16 to extract feature for loss