Why is the super-resolution result the same size as input image ?

[howfars]

I downloaded the repo as well as the pretrained model,trying to test the model.But I found that the size of the output image is the same as the input,and I think the resolution has no improvement.This is the result,the left image is original and the right one is result.

Here is my training settings,I only modified the file_list and gpu_mode option,other settings remain default.

# Training settings
parser = argparse.ArgumentParser(description='PyTorch Super Res Example')
parser.add_argument('--upscale_factor', type=int, default=4, help="super resolution upscale factor")
parser.add_argument('--testBatchSize', type=int, default=1, help='testing batch size')
# parser.add_argument('--gpu_mode', type=bool, default=True)
#Modify the gpu_mode variable to False,because I don't install cuda toolkit
parser.add_argument('--gpu_mode', type=bool, default=False)
parser.add_argument('--chop_forward', type=bool, default=False)
parser.add_argument('--threads', type=int, default=1, help='number of threads for data loader to use')
parser.add_argument('--seed', type=int, default=123, help='random seed to use. Default=123')
parser.add_argument('--gpus', default=1, type=int, help='number of gpu')
parser.add_argument('--data_dir', type=str, default='./Vid4')
parser.add_argument('--file_list', type=str, default='walk.txt')
parser.add_argument('--other_dataset', type=bool, default=True, help="use other dataset than vimeo-90k")
parser.add_argument('--future_frame', type=bool, default=True, help="use future frame")
parser.add_argument('--nFrames', type=int, default=7)
parser.add_argument('--model_type', type=str, default='RBPN')
parser.add_argument('--residual', type=bool, default=False)
parser.add_argument('--output', default='Results/', help='Location to save checkpoint models')
parser.add_argument('--model', default='weights/RBPN_4x.pth', help='sr pretrained base model')

Is there anything wrong?I'm a newbie in the super-resolution field,can someone help me?Thank you.

[YuanZYF]

I also encountered the same problem, how can I solve it?

[jsh-me]

I have the same problem
Somebody tell me the solution plz

[NabihGit]

anybody know the solution, I have the same problem

[NabihGit]

@YuanZYF @jsh-me @howfars So as what I find, u can just look at dataset.py , there the author down sample the src images to 1/4. u can just delete the downsample process (which is resize function in dataset.py), which include target, input and neibor images so the result image could be all right

[jsh-me]

@NabihGit
Oh! Thank you for letting me know. :)

[howfars]

@NabihGit Thank you so much!