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

Add doc for fast cycle detection #6230

Closed
ReyhaneAskari opened this issue Aug 1, 2017 · 3 comments · Fixed by #6000
Closed

Add doc for fast cycle detection #6230

ReyhaneAskari opened this issue Aug 1, 2017 · 3 comments · Fixed by #6000
Assignees
Milestone

Comments

@ReyhaneAskari
Copy link
Member

Regarding the improvement with the cycle_detection flag we need to add documentation both where all the flags are introduced and also we need to add some more explanations as notes in the doc so the users would be aware of the new feature.

@ReyhaneAskari ReyhaneAskari self-assigned this Aug 1, 2017
@lamblin lamblin added this to the 0.10 milestone Aug 1, 2017
@ReyhaneAskari
Copy link
Member Author

@lamblin , I checked and I have already included the explanation of cycle_detection in config.txt. Do you have any other suggestions on where I should add some side notes for it?

@nouiz
Copy link
Member

nouiz commented Aug 4, 2017

We need to let more people know about this. Few peole read completly config.txt and even then, don't fully understand all the consequence of each options. (or forget about them). So we must highligh that option more I think.

What about adding a note in this section: http://deeplearning.net/software/theano_versions/dev/tutorial/modes.html#optimizers to tell about this option that give a good compilation speed up for big graph? Tell that for big graph, it goes something between fast_compile and fast_run and that it is better to use that flag then optimizer_excluding=inplace.

@ReyhaneAskari
Copy link
Member Author

Sure. I will add it there.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

3 participants