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.

Sign up for GitHub

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

Jump to bottom

bandwidth parameter #35

Open
ADAM-CT opened this issue Jan 12, 2020 · 2 comments
Open

bandwidth parameter #35

ADAM-CT opened this issue Jan 12, 2020 · 2 comments

Comments

@ADAM-CT
Copy link

ADAM-CT commented Jan 12, 2020

First of all, I want to say that your work is amazing.
I used you in vgg16. gpus = 16. hybrid_conf.json. (https://github.com/msr-fiddle/pipedream/blob/f50827f2e28cbdbd82a4ea686c0498272b1460d6/runtime/image_classification/models/vgg16/gpus%3D16/hybrid_conf.json)

It only takes 600 seconds to train an epoch in the Imagenet dataset. Can you tell me how you generated this configuration file? Can you tell me the bandwidth parameter?
@deepakn94
Copy link
Collaborator

I'm not completely sure -- I think it was just the default value.

@ADAM-CT
Copy link
Author

ADAM-CT commented Jan 13, 2020

The default value is only one value (1000000000), only suitable for single server.I think our experimental environment is very similar, otherwise the performance will not be improved so much.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@deepakn94 @ADAM-CT