-
Notifications
You must be signed in to change notification settings - Fork 3.6k
updates teardown to account for ddp #2389
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
Conversation
| if self.global_rank == 0: | ||
| for proc in self.interactive_ddp_procs: | ||
| subprocess.Popen.kill(proc) |
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.
shouldn't it go into the run_training_teardown method?
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.
no because you still want certain things to happen with every process.
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.
oh wait, i see what you meant. i changed it. better now?
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.
not sure actually. I guess the "killing" should be the very last thing, after teardown has fully completed?
anyway not an expert here, so whatever works for you :)
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.
ah yes, i see now. That indeed should be the last thing since it will interrupt every other processes teardown. makes sense!
Codecov Report
@@ Coverage Diff @@
## master #2389 +/- ##
======================================
Coverage 88% 88%
======================================
Files 69 69
Lines 5452 5454 +2
======================================
+ Hits 4818 4820 +2
Misses 634 634 |
|
@awaelchli good to go with this? |
No description provided.