Cross-Scale-Non-Local-Attention icon indicating copy to clipboard operation
Cross-Scale-Non-Local-Attention copied to clipboard

Published 20 hours ago • verified publisher icon SHI-Labs

PSNR nan

Open JustinAsdz opened this issue 4 years ago • 1 comments

I create the LR dataset via resize function from opencv, using bicubic method. But when I start to train the model, the reported PSNR is always nan , both in X2 and X4 condition. Can you figure out where goes wrong? Thanks

image

JustinAsdz avatar Nov 30 '20 12:11 JustinAsdz

Did you change this line in option.py? previous: parser.add_argument('--data_range', type=str, default='1-800/800-810', help='train/test data range') after: parser.add_argument('--data_range', type=str, default='1-795/796-800', help='train/test data range')

I have encountered this problem. After changing this number, it was solved.

yuinnLIU avatar Dec 16 '20 07:12 yuinnLIU