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

Does MS1M Arcface or iBug exclude the overlapped data from LFW #566

Closed
lymanblue opened this issue Feb 27, 2019 · 6 comments
Closed

Does MS1M Arcface or iBug exclude the overlapped data from LFW #566

lymanblue opened this issue Feb 27, 2019 · 6 comments

Comments

@lymanblue
Copy link

Thank you for your great work!

Issue #24 shows that MS1M Arcface already remove the overlapped data from FaceScrub.
We would like to know the overlapped data of LFW is also removed or not.

Thank you.

@nttstar
Copy link
Collaborator

nttstar commented Mar 3, 2019

No.

@lymanblue
Copy link
Author

lymanblue commented Mar 4, 2019

From the experimental results in InsightFace, the 99.83% accuracy in LFW dataset is trained on MS1M-V2.

I am new to face recognition area.
I just wondering the training data MS1M-V2 (Arcface or IBUG?) in the paper removes the overlapped data with LFW or not.

Thank you.

@lymanblue
Copy link
Author

Hi~

Is dataset MS1M-IBUG the same with the cleaned MS1M dataset from the trillionpairs in Deep Glint?
The cleaned MS1M dataset states that they already exclude the overlapped data with LFW.

Thank you.

@nttstar
Copy link
Collaborator

nttstar commented Mar 4, 2019

Hi, the answer for both of your questions is No.

@lymanblue
Copy link
Author

Hi~

I originally thought MS1M-IBUG is the same with the cleaned MS1M dataset from Deep Glint.
It turns out that they are not the same.

So, what is the difference among MS1M-IBUG, MS1M-ArcFace and the cleaned MS1M?

Thank you.

@jetsmith
Copy link

@lymanblue did you figure out about the difference between these three datasets, is it possible to merge MS1M-ArcFace and glint dataset(including asian and msceleb-v1c)?

@nttstar nttstar closed this as completed Jun 1, 2023
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

3 participants