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

Benchmark tensorflow #65

Closed
cancan101 opened this issue Nov 9, 2015 · 9 comments
Closed

Benchmark tensorflow #65

cancan101 opened this issue Nov 9, 2015 · 9 comments

Comments

@cancan101
Copy link

Once there are examples of standard networks running on tensorflow.

@soumith
Copy link
Owner

soumith commented Nov 9, 2015

Thanks. On it right away!

@benanne
Copy link

benanne commented Nov 9, 2015

It seems to only support CuDNN v2, so I'm curious to see if it manages to be competitive.

@Jokeren
Copy link

Jokeren commented Nov 10, 2015

Very curious about its performance, as Jeff Dean hasn't mentioned much about its speed.

@adamist521
Copy link

Would this be same condition?
https://github.com/tensorflow/tensorflow/blob/master/tensorflow/models/image/alexnet/alexnet_benchmark.py

Done experiment on Tesla K-80 and showed around 180ms forward and 540ms forward-backward.

[edited out to avoid confusion]

@soumith
Copy link
Owner

soumith commented Nov 11, 2015

@adamist521 That benchmark is incomplete. It does not have FC layers, loss is different etc.
I should have numbers in about 1-2 hours. Finished AlexNet, Overfeat, working on VGG, Googlenet. Stay tuned.

@adamist521
Copy link

@soumith Awesome!

@bhack
Copy link

bhack commented Nov 11, 2015

👍

@bhack
Copy link

bhack commented Nov 11, 2015

@soumith take a look at tensorflow/tensorflow#113

@soumith
Copy link
Owner

soumith commented Nov 11, 2015

closing this. discuss here: #66

@soumith soumith closed this as completed Nov 11, 2015
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

6 participants