[dynamo] np.sort(complex) is not implemented #107710
Closed
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
This issue was discovered once we were able to trace without breaking in #107689. Same for the next one. [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/107710
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 900b37c with merge base ba5eeed ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Aug 22, 2023
lezcano
added
module: numpy
|
Funnily, related: |
ev-br
reviewed
Aug 22, 2023
There was a problem hiding this comment.
Optional: can also add same check to searchsorted.
I suppose we should either always return NotImplemented or always raise NotImplementedError? (Here and in normalize_not_implemented). As of this PR, using the order kwarg raises, but sorting a complex array returns.
| # `order` keyword arg is only relevant for structured dtypes; so not supported here. | ||
| a, axis, stable = _sort_helper(a, axis, kind, order) | ||
| result = torch.sort(a, dim=axis, stable=stable) | ||
| return result.values | ||
|
|
||
|
|
||
| def argsort(a: ArrayLike, axis=-1, kind=None, order: NotImplementedType = None): | ||
| if a.dtype.is_complex: | ||
| return NotImplemented |
There was a problem hiding this comment.
Not: Can move to _sort_helper to remove duplication.
There was a problem hiding this comment.
but would need to return somehow from the main function, so I didn't really know how to do it uniformly. I figured we could always just raise a NotImplementedError and then catch it in the normaliser and then return...
There was a problem hiding this comment.
Given that, a the moment, we always raise, we should probably go with the normaliser approach.
The point as to why doing the return rather than the exception is that dynamo already knows how to treat a function that returns NotImplemented, falling back gracefully to eager.
This issue was discovered once we were able to trace without breaking in #107689. Same for the next one. cc mruberry rgommers voznesenskym penguinwu anijain2305 EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx chenyang78 aakhundov [ghstack-poisoned]
|
Good point. Added searchsorted. |
ezyang
approved these changes
Aug 22, 2023
|
@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 |
This issue was discovered once we were able to trace without breaking in #107689. Same for the next one. cc mruberry rgommers voznesenskym penguinwu anijain2305 EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx chenyang78 aakhundov [ghstack-poisoned]
Merge failedReason: New commits were pushed while merging. Please rerun the merge command. Details for Dev Infra teamRaised by workflow job |
This issue was discovered once we were able to trace without breaking in #107689. Same for the next one. cc mruberry rgommers voznesenskym penguinwu anijain2305 EikanWang jgong5 Guobing-Chen XiaobingSuper zhuhaozhe blzheng Xia-Weiwen wenzhe-nrv jiayisunx chenyang78 aakhundov [ghstack-poisoned]
pytorchmergebot
pushed a commit
that referenced
this pull request
Aug 23, 2023
Gather does the same thing, but it's much better supported in the `torch.compile` stack Pull Request resolved: #107711 Approved by: https://github.com/ezyang ghstack dependencies: #107687, #107688, #107710
pytorchmergebot
pushed a commit
that referenced
this pull request
Aug 23, 2023
pytorchmergebot
pushed a commit
that referenced
this pull request
Sep 7, 2023
Follow-up to #107710: Help dynamo fall back to eager when compiling unimplemented numpy constructs: - arrays of strings - (arg){min, max} for complex types - various arguments typed as NotImplemented (`np.ones(4, order="F")` etc) - numpy functions which torch._numpy does not implement To test, run (we do not implement arrays of strings) ``` import torch import numpy as np @torch.compile(fullgraph=False) def fn(): return np.asarray(["L", "U"]) ``` and observe it compiles with fullgraph=False and fails with fullgraph=True Fixes #107970 Pull Request resolved: #107863 Approved by: https://github.com/ezyang, https://github.com/lezcano
ev-br
added a commit
to ev-br/pytorch
that referenced
this pull request
Oct 25, 2023
Follow-up to pytorch#107710: Help dynamo fall back to eager when compiling unimplemented numpy constructs: - arrays of strings - (arg){min, max} for complex types - various arguments typed as NotImplemented (`np.ones(4, order="F")` etc) - numpy functions which torch._numpy does not implement To test, run (we do not implement arrays of strings) ``` import torch import numpy as np @torch.compile(fullgraph=False) def fn(): return np.asarray(["L", "U"]) ``` and observe it compiles with fullgraph=False and fails with fullgraph=True Fixes pytorch#107970 Pull Request resolved: pytorch#107863 Approved by: https://github.com/ezyang, https://github.com/lezcano
ev-br
added a commit
to ev-br/pytorch
that referenced
this pull request
Oct 26, 2023
Follow-up to pytorch#107710: Help dynamo fall back to eager when compiling unimplemented numpy constructs: - arrays of strings - (arg){min, max} for complex types - various arguments typed as NotImplemented (`np.ones(4, order="F")` etc) - numpy functions which torch._numpy does not implement To test, run (we do not implement arrays of strings) ``` import torch import numpy as np @torch.compile(fullgraph=False) def fn(): return np.asarray(["L", "U"]) ``` and observe it compiles with fullgraph=False and fails with fullgraph=True Fixes pytorch#107970 Pull Request resolved: pytorch#107863 Approved by: https://github.com/ezyang, https://github.com/lezcano
ev-br
added a commit
to ev-br/pytorch
that referenced
this pull request
Oct 26, 2023
Follow-up to pytorch#107710: Help dynamo fall back to eager when compiling unimplemented numpy constructs: - arrays of strings - (arg){min, max} for complex types - various arguments typed as NotImplemented (`np.ones(4, order="F")` etc) - numpy functions which torch._numpy does not implement To test, run (we do not implement arrays of strings) ``` import torch import numpy as np @torch.compile(fullgraph=False) def fn(): return np.asarray(["L", "U"]) ``` and observe it compiles with fullgraph=False and fails with fullgraph=True Fixes pytorch#107970 Pull Request resolved: pytorch#107863 Approved by: https://github.com/ezyang, https://github.com/lezcano
ev-br
added a commit
to ev-br/pytorch
that referenced
this pull request
Oct 26, 2023
Follow-up to pytorch#107710: Help dynamo fall back to eager when compiling unimplemented numpy constructs: - arrays of strings - (arg){min, max} for complex types - various arguments typed as NotImplemented (`np.ones(4, order="F")` etc) - numpy functions which torch._numpy does not implement To test, run (we do not implement arrays of strings) ``` import torch import numpy as np @torch.compile(fullgraph=False) def fn(): return np.asarray(["L", "U"]) ``` and observe it compiles with fullgraph=False and fails with fullgraph=True Fixes pytorch#107970 Pull Request resolved: pytorch#107863 Approved by: https://github.com/ezyang, https://github.com/lezcano
ev-br
added a commit
to ev-br/pytorch
that referenced
this pull request
Oct 26, 2023
Follow-up to pytorch#107710: Help dynamo fall back to eager when compiling unimplemented numpy constructs: - arrays of strings - (arg){min, max} for complex types - various arguments typed as NotImplemented (`np.ones(4, order="F")` etc) - numpy functions which torch._numpy does not implement To test, run (we do not implement arrays of strings) ``` import torch import numpy as np @torch.compile(fullgraph=False) def fn(): return np.asarray(["L", "U"]) ``` and observe it compiles with fullgraph=False and fails with fullgraph=True Fixes pytorch#107970 Pull Request resolved: pytorch#107863 Approved by: https://github.com/ezyang, https://github.com/lezcano
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.
Stack from ghstack (oldest at bottom):
This issue was discovered once we were able to trace without breaking
in #107689. Same for the next
one.
cc @mruberry @rgommers @voznesenskym @penguinwu @anijain2305 @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @Xia-Weiwen @wenzhe-nrv @jiayisunx @chenyang78 @aakhundov