How to make DIV2K_train_LR_bicubic images ?

[BaeSukkyoung]

I want to train the EDSR with my own imageset.

Then, how can I make LR_bicubic images ?

Is there any code or it should be going by myself ?

[ulfailliyina]

try the code https://github.com/RakeshRaj97/downsampling-images-x2-super-resolution

[1632325673]

hello，i also train the EDSR with my own imageset,but I have encountered this problem, have you encountered it THCudaCheck FAIL file=/pytorch/aten/src/THC/THCGeneral.cpp line=405 error=11 : invalid argument [1600/16000] [L1: 17.1081] 6.4+47.0s [3200/16000] [L1: 11.4710] 6.6+7.4s [4800/16000] [L1: 9.1921] 7.5+9.3s [6400/16000] [L1: 7.8708] 6.4+6.4s [8000/16000] [L1: 7.0100] 8.1+9.7s [9600/16000] [L1: 6.4434] 7.3+7.8s [11200/16000] [L1: 6.0191] 8.5+12.8s [12800/16000] [L1: 5.6959] 9.4+14.7s [14400/16000] [L1: 5.4452] 7.3+10.3s [16000/16000] [L1: 5.2218] 7.2+9.8s

Evaluation: 0it [00:00, ?it/s] [div2k x2] PSNR: nan (Best: nan @epoch 1) Forward: 2.87s

Saving... Traceback (most recent call last): File "main.py", line 33, in main() File "main.py", line 28, in main t.test() File "/data2/qinfl/EDSR/src/trainer.py", line 123, in test self.ckp.save(self, epoch, is_best=(best[1][0, 0] + 1 == epoch)) File "/data2/qinfl/EDSR/src/utility.py", line 97, in save self.plot_psnr(epoch) File "/data2/qinfl/EDSR/src/utility.py", line 124, in plot_psnr label='Scale {}'.format(scale) File "/data2/qinfl/YES/envs/EDSR1/lib/python3.6/site-packages/matplotlib/pyplot.py", line 2842, in plot ({"data": data} if data is not None else {}), kwargs) File "/data2/qinfl/YES/envs/EDSR1/lib/python3.6/site-packages/matplotlib/axes/_axes.py", line 1743, in plot lines = [self._get_lines(args, data=data, **kwargs)] File "/data2/qinfl/YES/envs/EDSR1/lib/python3.6/site-packages/matplotlib/axes/_base.py", line 273, in call yield from self._plot_args(this, kwargs) File "/data2/qinfl/YES/envs/EDSR1/lib/python3.6/site-packages/matplotlib/axes/_base.py", line 399, in _plot_args raise ValueError(f"x and y must have same first dimension, but " ValueError: x and y must have same first dimension, but have shapes (0,) and (1,)