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
Feature: nadam optimizer #7715
Comments
I'm thinking about the same approach like |
Is there a reason this shouldn't simply be a community supplied option, like the cited version? Are you suggesting a specific modification to TF? Otherwise perhaps it would be better to discuss such an extension on stack overflow. |
@poxvoculi Do you think it makes sense to add a |
@ebrevdo: who would have an informed opinion on this? |
@poxvoculi A C++ implementation would be faster than python hacks like above. Also, maybe an awaiting googler label? |
Paul: probably vincent would know how popular / useful this optimizer is
for the general case.
…On Wed, Feb 22, 2017 at 11:53 AM, Aaron Hu ***@***.***> wrote:
@poxvoculi <https://github.com/poxvoculi> A C++ implementation would be
faster than python hacks like above. Also, maybe an awaiting googler label?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#7715 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ABtim1Y1n3pYI1s1-68zine8Pr6ees1Jks5rfJJHgaJpZM4MGufx>
.
|
Plausible. I have no experience with it. If someone implements it, it sounds like something we would incorporate. |
Automatically closing due to lack of recent activity. Please update the issue when new information becomes available, and we will reopen the issue. Thanks! |
Adam with nesterov momentum.
Keras has it: https://keras.io/optimizers/#nadam
TF implementation (old and doesn't quite work) : https://github.com/tdozat/Optimization/blob/master/tensorflow/nadam.py
The text was updated successfully, but these errors were encountered: