share apply_strategy method between autounit and autopredictunit
#512
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
Summary:
## Context:
Both `AutoUnit` and `AutoPredictUnit` use the same code block to apply the strategy on the module and check for any incompatibilties:
```
if strategy:
if isinstance(strategy, str):
strategy = _convert_str_to_strategy(strategy)
if isinstance(strategy, DDPStrategy):
if torch_compile_params and strategy.static_graph is True:
# https://dev-discuss.pytorch.org/t/torchdynamo-update-9-making-ddp-work-with-torchdynamo/860
raise RuntimeError(
"Torch compile requires DDPStrategy's static_graph to be False"
)
module = prepare_ddp(module, self.device, strategy)
elif isinstance(strategy, FSDPStrategy):
if swa_params:
raise RuntimeError(
"Stochastic Weight Averaging is currently not supported with the FSDP strategy"
)
# as stated here https://pytorch.org/get-started/pytorch-2.0/
rank_zero_warn(
"We recommend setting FSDPStrategy's use_original_params to True when using torch compile."
)
module = prepare_fsdp(module, self.device, strategy)
else:
module = module.to(self.device)
```
If changes are made to this logic, they must be made in both of those classes, which can be easily missed
## This Diff
Creates helper function `_apply_strategy_and_check(...)` to apply the strategy on the module and calls this function in both `AutoUnit` and `AutoPredictUnit` (other name suggestions are also welcome)
Differential Revision: D48612629
fbshipit-source-id: 4c43193b73a83ad5aaabe69582a213684421729f
|
This pull request was exported from Phabricator. Differential Revision: D48612629 |
Codecov Report
@@ Coverage Diff @@
## master #512 +/- ##
==========================================
+ Coverage 87.19% 87.30% +0.10%
==========================================
Files 106 106
Lines 8411 8403 -8
==========================================
+ Hits 7334 7336 +2
+ Misses 1077 1067 -10
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
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.
Summary:
Context:
Both
AutoUnitandAutoPredictUnituse the same code block to apply the strategy on the module and check for any incompatibilties:If changes are made to this logic, they must be made in both of those classes, which can be easily missed
This Diff
Creates helper function
_apply_strategy_and_check(...)to apply the strategy on the module and calls this function in bothAutoUnitandAutoPredictUnit(other name suggestions are also welcome)Differential Revision: D48612629