[benchmarks] Add scalar loss as model output when training#158074
Closed
benjaminglass1 wants to merge 11 commits intogh/benjaminglass1/93/basefrom
Closed
[benchmarks] Add scalar loss as model output when training#158074benjaminglass1 wants to merge 11 commits intogh/benjaminglass1/93/basefrom
benjaminglass1 wants to merge 11 commits intogh/benjaminglass1/93/basefrom
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/158074
Note: Links to docs will display an error until the docs builds have been completed. ❌ 6 New Failures, 1 Unrelated FailureAs of commit 1ec9923 with merge base 1abff80 ( NEW FAILURES - The following jobs have failed:
UNSTABLE - The following job is marked as unstable, possibly due to flakiness on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This was referenced Jul 10, 2025
Contributor
|
@anijain2305 , can you help to review this dynamo change? |
Collaborator
Author
|
@anijain2305 I ended up needing to add one new model XFAIL to this one, for |
Collaborator
Author
|
ROCm test failure appears unrelated to this PR. |
Collaborator
Author
|
Closing this, as we've decided not to pursue the AOTInductor training work at this time. |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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):
Add a hook to benchmark model forward passes that calculates a scalar loss as the first output when training (and detaches all other outputs). This is a requirement to use joint graph export (experimental), but can be done without loss of generality.
Additionally ensures that Dynamo traces through the loss calculation function (not done previously), which reduces the number of graph breaks in models when training.
cc @jeffdaily @sunway513 @jithunnair-amd @pruthvistony @ROCmSupport @dllehr-amd @jataylo @hongxiayang @naromero77amd @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @chenyang78 @kadeng @chauhang @amjames