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

M256 drop out rates in M1024 #88

Closed
khkk378 opened this issue May 20, 2021 · 2 comments
Closed

M256 drop out rates in M1024 #88

khkk378 opened this issue May 20, 2021 · 2 comments
Assignees

Comments

@khkk378
Copy link

khkk378 commented May 20, 2021

Hi!

You're using the M256 DO rates in M1024. Is this on purpose?

cdn1024 = Scaden(
        model_dir=model_dir + "/m1024",
        model_name="m1024",
        seed=seed,
        hidden_units=M1024_HIDDEN_UNITS,
        do_rates=M256_DO_RATES,
    )
@KevinMenden
Copy link
Owner

Hi Rasmus,

thanks for reporting! No that's not on purpose and a bug :) I refactored this part of the code recently, where I must have introduced this 🤦

Fortunately it doesn't matter, as this is the model used for prediction and dropout is deactivated during prediction. Nevertheless, not good. Will fix it!

@KevinMenden KevinMenden self-assigned this May 21, 2021
@KevinMenden
Copy link
Owner

Will probably get out a new release this evening because that's rather annoying.

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