fourson/deblurgan-pytorch

An easy-to-read implementation of DeblurGAN

deblurgan-pytorch's Issues

RuntimeError

File "I:\DeblurGAN-pytorch-master\train.py", line 80, in
main(config, args.resume)
File "I:\DeblurGAN-pytorch-master\train.py", line 54, in main
trainer.train()
File "I:\DeblurGAN-pytorch-master\base\base_trainer.py", line 100, in train
result = self._train_epoch(epoch)
File "I:\DeblurGAN-pytorch-master\trainer\trainer.py", line 152, in train_epoch
generator_loss.backward()
File "D:\miniconda3\lib\site-packages\torch_tensor.py", line 363, in backward
torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
File "D:\miniconda3\lib\site-packages\torch\autograd_init.py", line 173, in backward
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
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 3; expected version 2 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!

I try to run train.py, and I have this error. What should I do?

Save latest checkpoints

I want to create latest checkpoint after every epochs beside save period in config file. For example, I already set save period every 10 epochs but I still want to create latest checkpoint for every epoch. Can you guide me to do that? Thank you

_pickle.UnpicklingError: invalid load key, 'v'.

Thank you so much for the code, and open sourcing the weights.
When trying to use the checkpoint files for testing/deblurring
I am getting
"_pickle.UnpicklingError: invalid load key, 'v'."

Python Version : 3.6.3
Torch Version : 1.1.0

It seems to be a problem with the pickle library, what is the workaround?

pretrained weight file

I'm happy when i find the pytorch implementation of deblur gan .Could you please upload a copy of pretrained weight file since i want to to test the model quickly.Thank you .

RuntimeError:"_pickle.UnpicklingError: invalid load key, 'v'."

Hello! I have a pronlem when I run "python test.py --resume weights/WGAN_GP/checkpoint-epoch300.pth"or "python deblur_image.py --blurred data/blurred --deblurred data/sharp --resume weights/WGAN_GP/checkpoint-epoch300.pth" . It about "Traceback (most recent call last):
File "deblur_image.py", line 59, in
main(args.blurred, args.deblurred, args.resume)
File "deblur_image.py", line 12, in main
checkpoint = torch.load(resume)
File "/home/zongyb/.conda/envs/deblurGan/lib/python3.6/site-packages/torch/serialization.py", line 231, in load
return _load(f, map_location, pickle_module)
File "/home/zongyb/.conda/envs/deblurGan/lib/python3.6/site-packages/torch/serialization.py", line 369, in _load
magic_number = pickle_module.load(f)
_pickle.UnpicklingError: invalid load key, 'v'"
"

The problem of optimal weights for models

Hello! May I ask if you can provide the best weights for the model trained on the GOPRO dataset? Due to my limited computing power, I would like to use your weight directly as I will not use it for commercial purposes.

pre-train weights

Can you provide a loss curve corresponding to the pre-trained weights on the GoPro dataset? Thank you！

Cannot retrieve the pretrained models

HI fourson, thank you for your work!

However, I am encountering the issue of 'this repository is over its data quota' when I tried to pull the weights from git-lfs. I could not find the weights either in github.

Would you mind sharing it on google drive or somewhere else so that I could download it and share the joy of your work?

Thank you very much!

unrecognized arguments: weights\WGAN_GP

Hello. I try to run the program but I get an error. How can I solve that?

(base) C:\Users\EQUISPE\demo\DeblurGAN-pytorch-master>python deblur_image.py --blurred C:\Users\EQUISPE\demo\DeblurGAN-pytorch-master\prueba --deblurred C:\Users\EQUISPE\demo\DeblurGAN-pytorch-master\output --resume C:\Users\EQUISPE\demo\DeblurGAN-pytorch-master\checkpoints\pretrained weights\WGAN_GP
usage: deblur_image.py [-h] -b BLURRED -d DEBLURRED -r RESUME [--device DEVICE]

deblur_image.py: error: unrecognized arguments: weights\WGAN_GP