accelerator refactor - fix memory issue with ddp_spawn #5855
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
awaelchli
commented
Feb 6, 2021
| @@ -16,7 +16,6 @@ def setup(self, trainer, model): | |||
| raise MisconfigurationException(f"Device should be GPU, got {self.root_device} instead") | |||
| self.set_nvidia_flags() | |||
| torch.cuda.set_device(self.root_device) | |||
| model.to(self.root_device) | |||
| return super().setup(trainer, model) | |||
|
|
|||
There was a problem hiding this comment.
Just wanted to double check here with @tchaton and @justusschock
is it good to remove this? any implications for rpc/sequential?
For the other plugins I have not seen anything breaking.
justusschock
approved these changes
Feb 7, 2021
|
I'll close this in favor of #5866 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When using ddp spawn, moving the model to gpu before processes are spawn creates extra memory usage on the root gpu.
We already move the model to device in the plugins, so this shouldn't be necessary in setup.