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

The interest of minibatches #30

Open
Hebbalali opened this issue Oct 4, 2018 · 1 comment
Open

The interest of minibatches #30

Hebbalali opened this issue Oct 4, 2018 · 1 comment

Comments

@Hebbalali
Copy link

Hello,

I have a naive question about the use of minibatches. Since the ELBO you derive in your work is a generalization of the one used by Hensman,2013 (the not collapsed formulation) to multiple layers keeping the fully factorized form according to data points. The computation of the ELBO can be parallelised according to the data points.
Does minibatches allows this parallelization?

Thank you in advance for your answer

@hughsalimbeni
Copy link
Collaborator

Yes, you're exactly right. There is a parallelism over the data, so tensorflow is already computing things in parallel within the batch.

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