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

Quality Measure for GAN #12

Open
kingsj0405 opened this issue Feb 20, 2020 · 1 comment
Open

Quality Measure for GAN #12

kingsj0405 opened this issue Feb 20, 2020 · 1 comment
Milestone

Comments

@kingsj0405
Copy link
Member

PPT

  • TODO

내용

  • Inception Score(IS)
    • Entropy가 낮을수록 좋은 것이다
    • 다양한 Output을 가질수록 좋은 것이다
    • 이 두 개의 KL Divergency가 커야한다(롸?????)
    • 얘를 스코어로 줬을 때 딱히 잘 되진 않는다
    • 고양이를 넣었는데 강아지를 피크쳐도 상관없어 한다
    • 5만개 정도는 넣어야 신뢰성이 있다
  • Frechet Inception Distance(FID)
    • real data와 generating model data가 똑같기를 바라는 것이다
    • Inception Network의 중간을 본다
    • pre-trained 된 것을 사용하기 때문에 어느 정도 bias가 있다
  • Frechet Video Distance(FVD)
    • I3D를 써서 비디오를 한 게 있다
    • Gaussian은 너무 문제니까 Maximum Mean Discrepancy(MMD)를 쓰면 좋다
  • Q. Metric을 위한 Metric이 있나...?
  • A. 노가다 해야한다
@kingsj0405 kingsj0405 added this to the Week 5 milestone Feb 20, 2020
@miranheo
Copy link

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

2 participants