Fix 2.8 issue per sample grad #3460
Draft
+32
−15
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/pytorch/tutorials/3460
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New FailuresAs of commit d67bcb8 with merge base 9a44439 ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
svekars
commented
Jul 14, 2025
|
@albanD can you take a look? |
| print(f"Parameter {name}: max difference = {max_diff}") | ||
|
|
||
| # Optional: still assert for very large differences that might indicate real problems | ||
| assert max_diff < 0.5, f"Extremely large difference in {name}: {max_diff}" |
There was a problem hiding this comment.
Why did we change this to not use allclose anymore?
|
|
||
| # Print differences instead of asserting | ||
| max_diff = (per_sample_grad - ft_per_sample_grad).abs().max().item() | ||
| print(f"Parameter {name}: max difference = {max_diff}") |
There was a problem hiding this comment.
Why are we printing this? On a side note, could you share what this prints with the 2.8 RC?
Comment on lines
+175
to
+180
| idx = list(model.named_parameters()).index((name, model.get_parameter(name))) | ||
| per_sample_grad = per_sample_grads[idx] | ||
|
|
||
| # Check if shapes match and reshape if needed | ||
| if per_sample_grad.shape != ft_per_sample_grad.shape and per_sample_grad.numel() == ft_per_sample_grad.numel(): | ||
| ft_per_sample_grad = ft_per_sample_grad.view(per_sample_grad.shape) |
There was a problem hiding this comment.
I'm a bit confused by this part
Is the issue that torch.allclose now fails due to the ordering of per_sample_grad and ft_per_sample_grad being different during zip?
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.
Fixes https://github.com/pytorch/tutorials/actions/runs/16273680500/job/45947334370?pr=3416#step:9:2304