-
Notifications
You must be signed in to change notification settings - Fork 25.7k
[ONNX] Improve error handling for adaptive_pool #43032
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
…oof/fixAdaptiveAvg
💊 CI failures summary and remediationsAs of commit c8fc4bf (more details on the Dr. CI page): Commit c8fc4bf was recently pushed. Waiting for builds... This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group. This comment has been revised 52 times. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LG, thanks
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
…oof/fixAdaptiveAvg
|
@bzinodev Would you please let me know if this currently blocking merge? Or is there anything I can do to unblock this? |
Codecov Report
@@ Coverage Diff @@
## master #43032 +/- ##
==========================================
- Coverage 68.05% 68.04% -0.01%
==========================================
Files 396 396
Lines 51235 51238 +3
==========================================
Hits 34867 34867
- Misses 16368 16371 +3
Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
|
@bzinodev |
|
@bzinodev |
…oof/fixAdaptiveAvg
|
@bzinodev Can we merge this please? Thanks! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
…oof/fixAdaptiveAvg
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
|
cc @bzinodev |
| if mod != [0] * len(mod): | ||
| if output_size == [1] * len(output_size): | ||
| return g.op("GlobalMaxPool", input), None | ||
| return _unimplemented(name, 'output size that are not factor of input size') |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Old behavior a warning is emitted. It is changed to a runtime error? there is one internal test that fails. Is this something we want?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Note other ops all uses _unimplemented
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Before my update, this line in the symbolic conversion was throwing a warning instead of an error. Hence, the ATen node in the graph was not converted to ONNX, and was left as ATen.
By default, this behavior is not accepted. But exporter has a configuration mode that accepts combination of ONNX and ATen nodes. It's called ONNX_ATEN_FALLBACK.
It could be that the internal test case you see is failing is exporting with the ONNX_ATEN_FALLBACK mode.
I'll send a fix.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@bzinodev has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
|
@bzinodev Could you please merge the last commit here? There was a typo that was fixed after I addressed your comment. |
|
Hey, this PR unfortunately broke master (probably the missing typo you mentioned). I'm unlanding it, please resubmit the new PR with the fix, @bzinodev or myself can help you land it |
|
@bzinodev @dzhulgakov The new duplicate PR is: #45874 |
Summary: Duplicate of pytorch#43032 This update would also improve error handling for interpolate with 'area' mode. Pull Request resolved: pytorch#45874 Reviewed By: albanD Differential Revision: D24141266 Pulled By: bzinodev fbshipit-source-id: 7559f1d6af4f1ef3507c15a1aee76fe01fa433cd
This would also improve error handling for interpolate with 'area' mode.