search

sanghyun-son / EDSR-PyTorch

PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)
MIT License
2.42k stars 668 forks source link

training results good but the test PSNR results are 0? #240

Open Ccbov opened 4 years ago

Ccbov commented 4 years ago

image

why the test results PSNR are 0?

XuZikang commented 4 years ago

Hi, I've just encountered the same problem, have you solved it?

Ccbov commented 4 years ago

Hi, I've just encountered the same problem, have you solved it?

sorry, i dont

AbdulMoqeet commented 4 years ago

Are you using an updated EDSR code? I encountered the same error in the previous version but it has been resolved in the latest version.

dc-linkx commented 3 years ago

Hi, I've just encountered the same problem, have you solved it?

my test code : CUDA_VISIBLE_DEVICES=2,3 python main.py --data_test Demo --ext img --scale 2 --n_resblocks 32 --n_feats 256 --res_scale 0.1 --pre_train ../experiment/edsr_x2/model/model_best.pt --test_only --save_results

AbdulMoqeet commented 3 years ago

@dc-linkx Could you please attach a screenshot of output?

dc-linkx commented 3 years ago

@AbdulMoqeet here is the output , thx for your quick response image

AbdulMoqeet commented 3 years ago

Demo is just for super-resolving the images. You may calculate PSNR by either using Matlab Code or change -data_test Demo to -data_set my_dataset and keep my_dataset folder into benchmark folder with this kind of directory structure . Before calculating on your dataset, first try with benchmark dataset to ensure, you are doing right.

Liiiiaictx commented 7 months ago

@dc-linkx 大佬你解决了吗，我连进度条都显示不出来？

Liiiiaictx commented 7 months ago

@AbdulMoqeet 你好， 我用python main.py --data_test Demo --scale 2 --n_resblocks 32 --n_feats 256 --res_scale 0.1 --pre_train ../experiment/model/EDSR_x2.pt --test_only --save_results测试 出现以下几个问题 1