Tag functions to core IR in native_functions.yaml #105849
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/105849
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New FailuresAs of commit 866f3dd: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This PR needs a
|
|
This pull request was exported from Phabricator. Differential Revision: D47673670 |
|
This pull request was exported from Phabricator. Differential Revision: D47673670 |
|
FYI, you might also need to update test/expect/HasDecompTest.test_aten_core_operators.expect see #105673 for ref. |
|
This pull request was exported from Phabricator. Differential Revision: D47673670 |
|
This pull request was exported from Phabricator. Differential Revision: D47673670 |
There was a problem hiding this comment.
It seems not all tagged ops are documented in the spec https://pytorch.org/docs/stable/ir.html. Will the doc be updated accordingly? And does the "core IR" contain ops in both ATen IR and Prims IR?
| CompositeExplicitAutograd: logical_xor | ||
| tags: pointwise | ||
| tags: [core, pointwise] |
There was a problem hiding this comment.
Seems logical_xor is missing from the spec? https://pytorch.org/docs/stable/ir.html
There was a problem hiding this comment.
Yeah, so to clarify; that page is generated based on which ops are tagged as "core" from this YAML file. So once this is updated, it should be reflected in the page as well. The list of ops on the page is currently incomplete, and at the moment we are working on expanding the list. Hence this PR.
| @@ -1919,6 +1919,7 @@ | |||
| structured_delegate: cumsum.out | |||
| device_check: NoCheck # TensorIterator | |||
| variants: function, method | |||
| tags: core | |||
There was a problem hiding this comment.
cumsum is also not in the spec?
|
This pull request was exported from Phabricator. Differential Revision: D47673670 |
Summary: Pull Request resolved: pytorch#105849 Based on operator review meetings, tag appropriate functions as part of the Core IR. [Operator Review Tracking Sheet](https://docs.google.com/spreadsheets/d/1u9jQ-uGlKu-fe9nLy-jS2AIPtpE8sGTmELOFYgKOhXU/edit#gid=0) Test Plan: Use N3940835 to load the YAML and check updated core op list. Reviewed By: mergennachin, kimishpatel, SherlockNoMad Differential Revision: D47673670 fbshipit-source-id: e3b1afff0318451ce75d403a96844b0aee7d93a4
|
This pull request was exported from Phabricator. Differential Revision: D47673670 |
|
@pytorchbot merge (Initiating merge automatically since Phabricator Diff has merged) |
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 |
Summary: Based on operator review meetings, tag appropriate functions as part of the Core IR. [Operator Review Tracking Sheet](https://docs.google.com/spreadsheets/d/1u9jQ-uGlKu-fe9nLy-jS2AIPtpE8sGTmELOFYgKOhXU/edit#gid=0) Test Plan: Use N3940835 to load the YAML and check updated core op list. Reviewed By: mergennachin, kimishpatel, SherlockNoMad Differential Revision: D47673670 Pull Request resolved: pytorch#105849 Approved by: https://github.com/SherlockNoMad
Summary:
Based on operator review meetings, tag appropriate functions as part of the Core IR.
Operator Review Tracking Sheet
Test Plan: Use N3940835 to load the YAML and check updated core op list.
Reviewed By: mergennachin, kimishpatel, SherlockNoMad
Differential Revision: D47673670