search

sanghyun-son / EDSR-PyTorch

PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)
MIT License
2.43k stars 670 forks source link

error of weight dimensional #138

Open ch135 opened 5 years ago

ch135 commented 5 years ago

Hello, I am very grateful for your improved code.While, I have encountered the following problem when I run

python main.py --model EDSR --scale 2 --patch_size 96 --save edsr_baseline_x2 --reset --n_GPUs 2 --chop

Traceback (most recent call last): File "main.py", line 33, in main() File "main.py", line 28, in main t.test() File "/media/wangct/E7D0AC3987855C5C/ch_git/EDSR-PyTorch/src/trainer.py", line 89, in test sr = self.model(lr, idx_scale) File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call result = self.forward(*input, kwargs) File "/media/wangct/E7D0AC3987855C5C/ch_git/EDSR-PyTorch/src/model/init.py", line 57, in forward return forward_function(x) File "/media/wangct/E7D0AC3987855C5C/ch_git/EDSR-PyTorch/src/model/init.py", line 135, in forward_chop y = self.forward_chop(p, shave=shave, min_size=min_size) File "/media/wangct/E7D0AC3987855C5C/ch_git/EDSR-PyTorch/src/model/init.py", line 126, in forward_chop y = P.data_parallel(self.model, x, range(n_GPUs)) File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/parallel/data_parallel.py", line 188, in data_parallel outputs = parallel_apply(replicas, inputs, module_kwargs, used_device_ids) File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 83, in parallel_apply raise output File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/parallel/parallel_apply.py", line 59, in _worker output = module(*input, *kwargs) File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call result = self.forward(input, kwargs) File "/media/wangct/E7D0AC3987855C5C/ch_git/EDSR-PyTorch/src/model/edsr.py", line 52, in forward x = self.sub_mean(x) File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in call result = self.forward(*input, **kwargs) File "/home/wangct/.conda/envs/chenhao/lib/python3.6/site-packages/torch/nn/modules/conv.py", line 320, in forward self.padding, self.dilation, self.groups) RuntimeError: Expected 4-dimensional input for 4-dimensional weight [3, 3, 1, 1], but got 3-dimensional input of size [1, 184, 270] instead

Coucld you help me solv this problem. Looking forward to your busy reply！

sanghyun-son commented 5 years ago

Hello.

It seems that there are some problems with --chop argument #140 .

I will try to fix it.

Thank you!

ch135 commented 5 years ago

Hello.

It seems that there are some problems with --chop argument #140 .

I will try to fix it.

Thank you!

Execute me! Have you fix it？

ontheway16 commented 5 years ago

@thstkdgus35 Hi I am waiting for the same, I hope you may find some time to fix it soon, I was preparing to try it in another project, but failure of --chop stopped me badly. Thanks for the great code anyway, some tests surprisingly come out with "better then original" results, actually.

Raienryu97 commented 5 years ago

@ontheway16 could you please elaborate which tests and how? I'm curious and it might also help anyone in the future working on this too. Thanks.

fengshenfeilian commented 5 years ago

Excuse me, have you fix that problem?

sanghyun-son commented 5 years ago

@fengshenfeilian Hello. Would you specify your input? I cannot reproduce this error with my images.