SRGAN-tensorflow

SRGAN-tensorflow copied to clipboard

run error

1

InvalidArgumentError (see above for traceback): Cannot assign a device for operation 'dicriminator_train/discriminator/discriminator_unit/dense_layer_2/dense/bias/Adam_1': Operation was explicitly assigned to /device:GPU:0 but available devices are [ /job:localhost/replica:0/task:0/cpu:0 ]. Make sure the device specification...

lijain

Why is this repo's result better than the SRGAN paper?

3

I tried this repo and found that the result of SRGAN is better than in the paper. Not JUST better, it is abnormally better regarding the performance of GAN algorithm....

hsyoon94

KeyError: 'vgg19_1/vgg_19/conv5/conv5_4'

7

Traceback (most recent call last): File "main.py", line 231, in Net = SRGAN(data.inputs, data.targets, FLAGS) File "/home/zhou/SRGAN-tensorflow/lib/model.py", line 372, in SRGAN extracted_feature_gen = VGG19_slim(gen_output, FLAGS.perceptual_mode, reuse=False, scope=scope) File "/home/zhou/SRGAN-tensorflow/lib/model.py", line...

youthM

Module named "Keras" is not defined

11

I already installed "keras" and used import function in main.py but when I tried to train SRGAN with MSE this happened.

nickhdfan

scale generator network output to [-1, 1]

1

This generator network does not scale the output to [-1, 1] with tanh activation function. Does anyone know why?

pyguan88

why the content_loss and adversarial_loss keep increasing with the training?

2

3889187

Images not transforming correctly

13

For some reason images gets odd results when I try to alter them https://imgur.com/a/Q4wO3j9 Any ideas as to why? I use the pre-trained model with tensorflow CPU edition Edit: I...

Martan404

While running the pretrained model, how do we set the upscaling factor as 2 instead of the default which seems to be 4? Is this possible, if not, what's the...

snowPu

Missing requirements

1

Are able to provide a requirements.txt including the most suitable version of Tensorflow-gpu?

adamziek

Can I get output_image without target_image?

3

Hi. I'm Heidi. Can I get output_image without target_image? In main.py, the results are: `results = sess.run (save_fetch, feed_dict = {inputs_raw: input_im, targets_raw: target_im, path_LR: path_lr, path_HR: path_hr}` When I...

Heidisnaps