switch prefer_deferred_runtime_asserts_over_guards in export #160111
Closed
+86
−42
Conversation
This file contains hidden or 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
avikchaudhuri
requested review from
angelayi,
tugsbayasgalan,
ydwu4 and
zhxchen17
as code owners
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/160111
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ✅ No FailuresAs of commit 30e9d1e with merge base 4b4cdcf ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
5f14570 to
494f00b
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 7, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
494f00b to
d30a2b4
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 8, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 8, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
d30a2b4 to
a651671
Compare
pytorch-bot bot
pushed a commit
that referenced
this pull request
Aug 13, 2025
Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
a651671 to
55bc169
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 13, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
angelayi
reviewed
Aug 14, 2025
| dynamic_shapes: Optional[Union[dict[str, Any], tuple[Any, ...], list[Any]]] = None, | ||
| strict: bool = False, | ||
| preserve_module_call_signature: tuple[str, ...] = (), | ||
| prefer_deferred_runtime_asserts_over_guards: bool = False, |
There was a problem hiding this comment.
should this be in the toplevel export API?
There was a problem hiding this comment.
I feel you also need to propagate it to export_for_training as well.
There was a problem hiding this comment.
@angelayi Yes, intentional: I pushed it up to the top level to act as a switch to manage risk.
@tugsbayasgalan The tests didn't need me to push this to export_for_training. Since it's on the way to deprecation, leave as is?
There was a problem hiding this comment.
Hmm interesting. Today torch.export.export calls into torch.export._trace._export_for_training at pre-dispatch level. So if tests are not catching it, something seems off. I still think we should add it tho because the way we are marketing the deprecation is that torch.export.export and torch.export.export_for_training are exact same APIs. Having different flags will make them look different.
There was a problem hiding this comment.
To be clear, _export_for_training does have this flag. Anyway, I can add it.
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
55bc169 to
c841051
Compare
pytorch-bot bot
pushed a commit
that referenced
this pull request
Aug 21, 2025
Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 21, 2025
…#160111) Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
c841051 to
1b09fe1
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 21, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 21, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 26, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
force-pushed
the
export-D79734206
branch
2 times, most recently
from
0b17893 to
2014a9b
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 26, 2025
…#160111) Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 26, 2025
…#160111) Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
2014a9b to
d87ccec
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 26, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 27, 2025
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
d87ccec to
f5a6f07
Compare
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 27, 2025
…#160111) Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
avikchaudhuri
added a commit
to avikchaudhuri/pytorch
that referenced
this pull request
Aug 27, 2025
…#160111) Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
…#160111) Summary: Pull Request resolved: pytorch#160111 In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206
|
This pull request was exported from Phabricator. Differential Revision: D79734206 |
avikchaudhuri
force-pushed
the
export-D79734206
branch
from
f5a6f07 to
30e9d1e
Compare
tugsbayasgalan
approved these changes
Aug 27, 2025
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
|
@pytorchbot merge (Initiating merge automatically since Phabricator Diff has merged) |
|
Can't merge closed PR #160111 |
markc-614
pushed a commit
to markc-614/pytorch
that referenced
this pull request
Sep 17, 2025
…#160111) Summary: In preparation for checking shape guards in export, this PR effectively switches `prefer_deferred_runtime_asserts_over_guards` to `False`, matching Dynamo. Actually that's a lie: we switch it to `allow_complex_guards_as_runtime_asserts`, which is `False` by default but can be controlled via an internally API to be `True`. This makes the two flags synchronized, so we should be able to kill `allow_complex_guards_as_runtime_asserts` at this point. Test Plan: updated tests Rollback Plan: Differential Revision: D79734206 Pull Request resolved: pytorch#160111 Approved by: https://github.com/tugsbayasgalan
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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:
In preparation for checking shape guards in export, this PR effectively switches
prefer_deferred_runtime_asserts_over_guardstoFalse, matching Dynamo.Actually that's a lie: we switch it to
allow_complex_guards_as_runtime_asserts, which isFalseby default but can be controlled via an internally API to beTrue. This makes the two flags synchronized, so we should be able to killallow_complex_guards_as_runtime_assertsat this point.Test Plan:
updated tests
Rollback Plan:
Differential Revision: D79734206
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben