Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Set5 test set evaluation result #16

Open
LiUzHiAn opened this issue Feb 25, 2019 · 6 comments
Open

Set5 test set evaluation result #16

LiUzHiAn opened this issue Feb 25, 2019 · 6 comments

Comments

@LiUzHiAn
Copy link

Hi,
I used the model you pretrained with 200 epochs and evaluated the model on Set5 test set. But I only got ~25.5 average PSNR which is somewhat poor. Anybody same with me? Should I train the model continuely ?

@13135675192
Copy link

hello!How long have you been training this network?

@LiUzHiAn
Copy link
Author

LiUzHiAn commented Mar 5, 2019

I used Set91 and BSD200 as training set. Traning data 200 epochs on Colab nearly cost 5 hrs

@13135675192
Copy link

I used DIV2K as training set. and set "matlab -true "on GTX-1080ti nearly cost 8days.
The result is 34.5average PSNR,the testing set is set5

@LiUzHiAn
Copy link
Author

oh, the result is not that good even worse than SRCNN. Have you got any improvement ?

@sxd0071
Copy link

sxd0071 commented Mar 11, 2019

@13135675192 Hey, how many training images do you have. After preprocessing the traing images, I got about 500 thousand images. When I train this network with scale of 3 for seven days , the trainning step is just 800 thousand.

@kong1060553768
Copy link

@ 13135675192 May I ask you 34.5 average PSNR at scale3 or 2?

Can you use your checkpoint folder for reference? My mailbox is 1060553768@qq.com

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants