the size is wrong

[Winnie202]

Thank you very much for your code contribution, after I run according to your code,like this:

the test data shape is reds, lr image shape is 3232, hr image shape is 128128, but an error occurred,like this:

I try to modify some parameters,in model/CRFP.py:

but an error occurred,like this:

Excuse me, where did I go wrong

[woctezuma]

Would it make sense to use the following argument?

--GT_size 128

in order not use the default value of 256:

CRFP/option.py

Lines 88 to 89 in c7d0b82

	parser.add_argument('--GT_size', type=int, default=256,
	help='Training GT size')

[Winnie202]

Would it make sense to use the following argument?

--GT_size 128

in order not use the default value of 256:

CRFP/option.py

Lines 88 to 89 in c7d0b82

	parser.add_argument('--GT_size', type=int, default=256,
	help='Training GT size')

No, after the parameters are modified:--GT_size 128, an error is still reported

[lienfeng1011]

Would it make sense to use the following argument?

--GT_size 128

in order not use the default value of 256:

CRFP/option.py

Lines 88 to 89 in c7d0b82

	parser.add_argument('--GT_size', type=int, default=256,
	help='Training GT size')

No, after the parameters are modified:--GT_size 128, an error is still reported

Hi, could you please check TestSet() class in dataset/reds.py?
Since we source the different preprocessed LR dataset with different scaling factor.
And in test mode we use full size of data and the parameters --GT_size and --LR_size will not be used.
The size could be wrong at the begining when you load those data.

And I also wonder what are the shapes of these tensors in your case? Thank you.