[ONNX][dynamo_export] Turn off opmath for type promotion #113780
Closed
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
[ghstack-poisoned]
pytorch-bot
bot
added
the
release notes: onnx
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/113780
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit 726d6bc with merge base aeb5fd5 ( FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
[ghstack-poisoned]
BowenBao
requested review from
abock,
thiagocrepaldi and
wschin
as code owners
titaiwangms
reviewed
Nov 15, 2023
justinchuby
reviewed
Nov 15, 2023
[ghstack-poisoned]
… for type promotion" [ghstack-poisoned]
justinchuby
reviewed
Nov 16, 2023
| "native_batch_norm", | ||
| "dot", | ||
| "logit", | ||
| "rsub", |
There was a problem hiding this comment.
At some point we can refactor this file to follow what’s in ONNX Script so the meta data is clearer
…ime' on "[ONNX][dynamo_export] Turn off opmath for type promotion" [ghstack-poisoned]
[ghstack-poisoned]
titaiwangms
approved these changes
Nov 22, 2023
justinchuby
approved these changes
Nov 22, 2023
titaiwangms
reviewed
Nov 22, 2023
| "nn.functional.celu", | ||
| dtypes=(torch.float16,), | ||
| reason=onnx_test_common.reason_onnx_does_not_support("Celu", "float16"), | ||
| ), |
There was a problem hiding this comment.
non-blocking microsoft/onnxscript#1158
pytorchmergebot
pushed a commit
that referenced
this pull request
Nov 22, 2023
Such that even if failures occur during model run, the sample inputs are accessible for later investigation. Pull Request resolved: #114163 Approved by: https://github.com/thiagocrepaldi ghstack dependencies: #113780
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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):
Although opmath is the right thing to do to retain on-par precision, it inserts
upcasts everywhere in the graph. This is particularly hard for backend to optimize
since there is no way to differentiate between inserted upcasts and model code
casts. Hence we consolidate the input dtype to the result dtype to avoid this.