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

Low SimCLR basseline #1

Open
TsungWeiTsai opened this issue Dec 29, 2020 · 0 comments
Open

Low SimCLR basseline #1

TsungWeiTsai opened this issue Dec 29, 2020 · 0 comments

Comments

@TsungWeiTsai
Copy link

TsungWeiTsai commented Dec 29, 2020

Hi author of CLAE,

Thank you for the insightful paper. It would be great if you plan to release the code in the near future so that I can have a better understanding of your work.

Besides, from your paper, the performance of SimCLR is much lower than the ones reported in the original paper (90+% accuracy on CIFAR-10 training from scratch). I am wondering if there is any reason that you are using a much weaker baseline. Since SimCLR also reported the hyper-parameters they used, it would be interesting to see if CLAE can still improve over the strong baselines. Thank you for your time in advance.

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

1 participant