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

How to test? Please!!! I have got the pre-trained models， but i don't know the test commend。 #300

Open Studyhao opened 3 years ago

Awenjie10 commented 3 years ago

hello, Have you learned about it? I got some mistakes about the testing, I used my own images, but the model outputs only artifacts.

mylifeasazucchini commented 3 years ago

Maybe I am too late but for anyone else, you need to add a few additional parameters and the function call should look like this:

python main.py --data_test Demo --scale 4 --pre_train ../models/EDSR_x4.pt --n_resblocks 32 --n_feats 256 --res_scale 0.1 --test_only --save_results

This is after creating a folder containing your test images called 'test' in same working directory where you have other folders like src, models, etc.

haikunzhang95 commented 2 years ago

Thank you very much! It's works for me.

ymtupup commented 2 years ago

Why the --save_results come out error? 捕获1

ambikaverma commented 2 years ago

I'm getting same error as above for --save_result. any recommendations? @mylifeasazucchini

update - https://github.com/sanghyun-son/EDSR-PyTorch/issues/320#issuecomment-936413370

Major update - working now. I had missed adding --n_resblocks 32 and --res_scale 0. now its working as expected. thank you!

Ralapfranco commented 2 years ago

n_resblocks 32 --n_feats 256 --res_scale 0.1 which i use" n_resblocks 32 --n_feats 256 --res_scale 0.1"it is worse than that which i do not use the“ n_resblocks 32 --n_feats 256 --res_scale 0.1”, when i test x3_moedl.

liu1216 commented 1 year ago

hello, Have you learned about it? I got some mistakes about the testing, I used my own images, but the model outputs only artifacts.

Excuse me, have you solved it? When I test with my own test set, I can only output super-resolution images, and PSNR = 0?

Ralapfranco commented 1 year ago

sorry! i'm too late. You should debug it！ i guess that the module can't get the test path.

At 2022-08-30 16:39:18, "liu1216" @.***> wrote:

hello, Have you learned about it? I got some mistakes about the testing, I used my own images, but the model outputs only artifacts.

Excuse me, have you solved it? When I test with my own test set, I can only output super-resolution images, and PSNR = 0?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>