[export] hook up mark_dynamic to export Dims #137029
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/137029
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (2 Unrelated Failures)As of commit 7913b67 with merge base 475a8a4 ( FLAKY - The following jobs failed but were likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
@pianpwk has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
torch/export/dynamic_shapes.py
Outdated
| import sympy | ||
|
|
||
| from torch.fx.experimental.symbolic_shapes import StrictMinMaxConstraint | ||
| from torch.fx.experimental.symbolic_shapes import RelaxedUnspecConstraint, StrictMinMaxConstraint |
There was a problem hiding this comment.
| from torch.fx.experimental.symbolic_shapes import RelaxedUnspecConstraint, StrictMinMaxConstraint | |
| from torch.fx.experimental.symbolic_shapes import StrictMinMaxConstraint |
pianpwk
requested review from
angelayi,
avikchaudhuri,
tugsbayasgalan,
ydwu4 and
zhxchen17
as code owners
|
@pianpwk has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
|
||
| with self.assertRaisesRegex( | ||
| torch._dynamo.exc.UserError, | ||
| r"Not all values of RelaxedUnspecConstraint.* are valid because .* was inferred to be a constant", |
There was a problem hiding this comment.
nit: we shouldn't be exposing internal names like RelaxedUnspecConstraint in user-facing errors, they're noise.
|
@pytorchbot merge -f 'Landed internally' (Initiating merge automatically since Phabricator Diff has merged, using force because this PR might not pass merge_rules.json but landed internally) |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Please use Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Adds Dim.DYNAMIC which calls torch._dynamo.mark_dynamic() in the backend. Similar to Dim.AUTO in that it does automatic inference for ranges & relations, but errors out for specializations. Pull Request resolved: pytorch#137029 Approved by: https://github.com/avikchaudhuri
4 participants
Adds Dim.DYNAMIC which calls torch._dynamo.mark_dynamic() in the backend. Similar to Dim.AUTO in that it does automatic inference for ranges & relations, but errors out for specializations.