Skip to content

Conversation

karmel
Copy link
Contributor

@karmel karmel commented Mar 20, 2018

This gives us an extra 0.5% - 1.5% examples/second, depending on the environment, and should not in theory change the model architecture itself. Final accuracy for ImageNet was unchanged at 76.38%, CIFAR-10 92.5%.

CC @bignamehyp @reedwm

@karmel karmel requested review from nealwu and robieta March 20, 2018 21:40
@karmel karmel requested a review from k-w-w as a code owner March 20, 2018 21:40
@k-w-w
Copy link
Contributor

k-w-w commented Mar 20, 2018

Nice!
Can you rename final_avg_pool to final_reduce_mean in imagenet_test.py as well?

@karmel
Copy link
Contributor Author

karmel commented Mar 20, 2018

Nice catch. Tests fixed. (Linting being fixed in #3676 .)

Copy link
Contributor

@k-w-w k-w-w left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

looks good to me!

@karmel karmel merged commit d5663b3 into master Mar 20, 2018
@karmel karmel deleted the fix/perf-tune-model branch March 20, 2018 23:13
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants