ayu-22 / bppnet-back-projected-pyramid-network Goto Github PK

NameError                                 Traceback (most recent call last)
<ipython-input-26-a24d9fb97dd3> in <module>
      4 input_dis_channel = 3
      5 max_epochs = 100
----> 6 DUNet = DU_Net(input_unet_channel ,output_unet_channel ,input_dis_channel).cuda()

<ipython-input-23-e5148fdf2410> in __init__(self, unet_input, unet_output, discriminator_input)
     27         self.add_module('l1_loss', l1_loss)
     28         self.add_module('content_loss', content_loss)
---> 29         self.add_module('style_loss', style_loss)
     30         self.add_module('ssim_loss', ssim)
     31         self.add_module('bce_loss', bce)

NameError: name 'style_loss' is not defined

As the error message said, style_loss is not defined in the code

I am using the weights that were emailed to me.
This is the kind of output that I am getting.
Before:

After:

I can't seem to figure out what is wrong. Any idea what might be wrong?

In function process, the second return value should logically be DIS_LOSS, but in your code it is the SSIM value. The function calculated DIS_LOSS but did not use it. What is the purpose of this?I tried to change the second return value to DIS_LOSS, but there is a problem that DIS_LOSS will not converge.

Sorry but i can't get your pytorch version.
Meanwhile, some other packages in the requirements.txt don't seem necessary.
New to this, plz ignore me if I'm asking silly.

I would like to ask, the official I-HAZE dataset states that there are 35 pairs of images, but after downloading from the official website, there are only 30 pairs. May I ask where did you obtain the complete I-HAZE dataset of 35 pairs of images?

If possible, could you please send me the complete I-HAZE dataset via email? I would greatly appreciate it? My email is [email protected]

Hello,Can you issue the code of the model?

I tried to train with Ohaze dataset, with batch len=210 (default setting), 35 images for training, 5 for test, i have 2 problems that need some help, first, which images are your test ones ?except 2 images shown on the gif i selected , which 3 left are your testing choices? second, except 5 for test 35 for training, there are still 5 left for validation, but i cant find any validation codes in file, should i put these 5 images in training? for i cant get that high psnr and ssim as paper declaimed, specifically, psnr 19.1+- with ssim 0.75+, thank you

Hi, when I use dataset O-Haze and I-Haze to run "train(epochs)", the following error message appears:

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]], which is output 4 of BroadcastBackward, is at version 2; expected version 1 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

What is the cause？ Thx

I would appreciate it if you could send the pre-trained weights to me. I sent you an email but haven't got any responses.

.

RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [1, 512, 4, 4]] is at version 2; expected version 1 instead. Hint: the backtrace further above shows the operation that failed to compute its gradient. The variable in question was changed in there or anywhere later. Good luck!

Running into this error, while training

1.I want to ask which loss of MSE, SSIM, Unet loss is the standard? And what conditions are stable?
Whenever the picture is improved, the learning rate will decrease,
but the picture-like viewing will get worse.
2.This is my picture result. How can I improve it? Thank you.

Hi is it possible to obtain the pre-trained weights?