-
Notifications
You must be signed in to change notification settings - Fork 3.4k
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
Fix double precision + ddp_spawn #6924
Conversation
Codecov Report
@@ Coverage Diff @@
## master #6924 +/- ##
=======================================
+ Coverage 88% 92% +4%
=======================================
Files 199 199
Lines 12987 13008 +21
=======================================
+ Hits 11412 12006 +594
+ Misses 1575 1002 -573 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can you reconsider this PR with respect to concurrent #7289?
This pull request has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. If you need further help see our docs: https://pytorch-lightning.readthedocs.io/en/latest/generated/CONTRIBUTING.html#pull-request or ask the assistance of a core contributor here or on Slack. Thank you for your contributions. |
@Borda I don't think we can use |
@Borda @awaelchli @justusschock I've updated the PR, mind re-reviewing? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM !
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Note the test for double precision now takes 70+ seconds because of spawn getting called six times 2 x (fit, test, predict)
This is not ideal especially for a test that does not assert anything.
@awaelchli I've simplified the test so now should only spawn once 😃 |
What does this PR do?
Fixes #6893 - should also prevent issue from discussion #6851
Basically just changes patching to wrapping, had to add a new base wrapper class and support for unwrapping, but it means that double precision + distributed wrappers works fine and no longer needs a teardown.
Before submitting
PR review
Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:
Did you have fun?
Make sure you had fun coding 🙃