[dynamo] Account for function id reuse in relevant Dynamo decorators #148385
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
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/148385
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit f6ee148 with merge base aade4fb ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Mar 3, 2025
zou3519
reviewed
Mar 4, 2025
torch/_dynamo/decorators.py
Outdated
Comment on lines
165
to
166
| trace_rules._allowed_callable_ids.remove(fn_id) | ||
| trace_rules._nonstrict_trace_callable_ids.remove(fn_id) |
There was a problem hiding this comment.
do we need to remove stuff from _disallowed_callable_ids?
There was a problem hiding this comment.
OOps copy paste error.
Removed trace_rules._nonstrict_trace_callable_ids.remove(fn_id), and we don't need to touch _disallowed_callable_ids because allow_in_graph was removing fn_id from it, so it's safe.
|
Starting merge as part of PR stack under #148007 |
pytorchmergebot
pushed a commit
that referenced
this pull request
Mar 5, 2025
…ant`-ed instances (#148007) As title, this enables `nonstrict_trace`-ed function to take in object whose type has been `pytree.register_constant`-ed, as long as the object existed outside the `torch.compile` region. This also forces Dynamo to emit a `EQUALS_MATCH` guard on the object. Pull Request resolved: #148007 Approved by: https://github.com/zou3519 ghstack dependencies: #148385
pytorchmergebot
pushed a commit
that referenced
this pull request
Mar 5, 2025
This fixes a recent series of flaky failure from `nonstrict_trace` unit tests: #148166, #148056, #148055, #148054, #148034, #148033, #148032, #148031. For now we don't need to worry about the other decorators because they are either meant for builtin/numpy functions (which should never deallocate in practice), or used for polyfills which keeps the function object in `get_torch_obj_rule_map()`. Fixes #147777. ghstack-source-id: d9bea5f Pull Request resolved: #148385
This was referenced Mar 10, 2025
Closed
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):
functools.partialobjects #148386nonstrict_tracework with somepytree.register_constant-ed instances #148007This fixes a recent series of flaky failure from
nonstrict_traceunittests: #148166, #148056, #148055, #148054, #148034, #148033, #148032, #148031.
For now we don't need to worry about the other decorators because they
are either meant for builtin/numpy functions (which should never
deallocate in practice), or used for polyfills which keeps the function
object in
get_torch_obj_rule_map().Fixes #147777.
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang @amjames