Throw error when torch iterable was not split by rank or split by worker #107
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
AlirezaSohofi
suggested changes
Jan 18, 2023
| yield from self.source | ||
|
|
||
| def _check_for_rank_split(self, source: Composable) -> bool: |
There was a problem hiding this comment.
here we should check if we are actually in the multi-rank environment. If not this check should be skipped.
Would be good to add a test for this case too: we are not in the multi-rank / multi-worker environment and to_torch_iterable is called.
| else: | ||
| return self._check_for_rank_split(source.source) | ||
|
|
||
| def _check_for_worker_split(self, source: Composable) -> bool: |
| @@ -66,8 +66,36 @@ def __init__(self) -> None: | |||
|
|
|||
| def __iter__(self) -> Iterator: | |||
| """Method to iterate over the source""" | |||
| if not self._check_for_rank_split(self.source): | |||
| raise ValueError( | |||
There was a problem hiding this comment.
We can have a custom exception for this, PytorchSplittingError
| .split_by_worker_pytorch() | ||
| .to_torch_iterable() | ||
| ) | ||
| next(iter(it)) |
There was a problem hiding this comment.
Please check the length or items that we get.
pzdkn
force-pushed
the
peng-throw-error-for-omitting-split-in-torch-mp
branch
from
48462e9
to
c028ecc
Compare
AlirezaSohofi
previously approved these changes
Feb 2, 2023
| @@ -66,8 +68,40 @@ def __init__(self) -> None: | |||
|
|
|||
| def __iter__(self) -> Iterator: | |||
| """Method to iterate over the source""" | |||
| if _in_multi_rank_env(): | |||
| warnings.warn("In multi rank environment") | |||
There was a problem hiding this comment.
We don't need warning here
| "Add a 'split_by_rank_pytorch' call to your composable to avoid this error. " | ||
| ) | ||
| if _in_multi_worker_env(): | ||
| warnings.warn("In mulit worker environment") |
There was a problem hiding this comment.
We don't need warning here
| if torch.distributed.is_available() and torch.distributed.is_initialized(): | ||
| group = torch.distributed.group.WORLD | ||
| size = torch.distributed.get_world_size(group=group) | ||
| return True if torch.distributed.is_available() and size > 1 else False |
There was a problem hiding this comment.
torch.distributed.is_available() is already checked in line 139
pzdkn
force-pushed
the
peng-throw-error-for-omitting-split-in-torch-mp
branch
from
3936284
to
ccfcdc8
Compare
AlirezaSohofi
previously approved these changes
Feb 10, 2023
AlirezaSohofi
approved these changes
Feb 10, 2023
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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.
Description
Please include a summary of the change and which issue is fixed. Please also include relevant motivation and context.
List any dependencies that are required for this change.
Fixes # issue
Type of change
Checklist: