-
-
Notifications
You must be signed in to change notification settings - Fork 611
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
Optimistic ADAM #1246
Optimistic ADAM #1246
Conversation
are sure about the implementation? it seems different from the one in the paper |
actually I was wrong, the implementation is correct |
Thanks! bors r+ |
pff, this news thing produces a lot of conflicts the merging PRs one after another |
Merge conflict. |
should I do a rebase or something? |
Yes, rebase please
Il mar 30 giu 2020, 20:23 cossio <notifications@github.com> ha scritto:
… should I do a rebase or something?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
<#1246 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABVQOQSQ424CWFSQPDDAKGDRZIUQLANCNFSM4OE3F3TQ>
.
|
Rebased |
Is the CI fail on Julia 1 important? Doesn't look related to this PR. |
the failure is on the current stable version but it's unrelated to this pr (is a doctest fail) bors r+ |
Build succeeded: |
Optimistic ADAM is a variant of ADAM which adds an "optimistic" term, suitable for adversarial training of a model. The optimistic term can be interpreted as predicting the gradient of the adversary.
Reference: https://par.nsf.gov/biblio/10079723-training-gans-optimism
PR Checklist
@MikeInnes
or@dhairyagandhi96
(for API changes).