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

FID 和LPIPS #3

Open
ludandandan opened this issue Jan 28, 2024 · 6 comments
Open

FID 和LPIPS #3

ludandandan opened this issue Jan 28, 2024 · 6 comments

Comments

@ludandandan
Copy link

你好,使用pre-trained model
和dataset FS2K生成的结果(1037张素描图【由于test的photo只有1037张,这也是一个疑问点】)计算FID,分别使用了https://github.com/AiArt-HDU/HIDA/blob/main/evaluation/dirs_fid_score.py和https://github.com/mseitzer/pytorch-fid的方法计算的FID得分都是58左右(ground truth是dataset FS2K中的sketch),并且对FSGAN计算的FID(groud truth使用的是FS2K原版的图像,由于FS2K中图像尺寸不一,全部resize到256x256计算的)得分在44左右,使用https://github.com/richzhang/PerceptualSimilarity计算的LPIPS,HIDA方法得分是0.273(这里只有1037张图),而FSGAN计算的是0.228(这里是正常的1046张图),除了0.273这个数值外,其他数值与论文中的差距有些大,请问是我哪里处理不当吗?

@zjyruobing
Copy link

请问你们可以跑通训练的代码吗

@ludandandan
Copy link
Author

我试了一下测试的代码,可以生成图像,训练的还没搞

@JcccKing
Copy link
Contributor

The size of the image generated may be 250 * 250

@Lun-menG
Copy link

你好,请问你们的anno_train.json文件是在哪里获取的呢

@ludandandan
Copy link
Author

@Lun-menG
Copy link

https://github.com/DengPingFan/FS2K从这里下载的
哦哦,好的,非常感谢

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