[pt2e] Fix QAT annotations for special qspecs #3337
Merged
+159
−2
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/ao/3337
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit c19b8cd with merge base e2aab90 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
@andrewor14 has imported this pull request. If you are a Meta employee, you can view this in D86983409. |
jerryzh168
reviewed
Nov 13, 2025
Contributor
jerryzh168
left a comment
There was a problem hiding this comment.
ah makes sense, can you add a test
Contributor
Author
Yep, this is in progress, just wanted to push first for Xing to try |
646cd49 to
d76110c
Compare
jerryzh168
reviewed
Nov 13, 2025
| for n in nodes_to_check: | ||
| if n.target == torch.ops.aten.batch_norm.default: | ||
| num_batch_norm_nodes_checked += 1 | ||
| self.assertTrue(n not in old_nodes, "found old node in qspec") |
Contributor
There was a problem hiding this comment.
is it possible to also check they are all in new_nodes?
jerryzh168
approved these changes
Nov 13, 2025
Contributor
jerryzh168
left a comment
There was a problem hiding this comment.
lg, thanks for the fix, had one inline comment
**Summary:** In pt2e QAT, we first annotate the nodes to be quantized and then perform the pattern replacement (i.e. QAT fusion) in the prepare step. After this pattern replacement, old nodes are replaced with new nodes, and any references to the old nodes must also be updated to refer to the new nodes. This commit fixes a bug where, for special qspecs like the `SharedQuantizationSpec` and `DerivedQuantizationSpec`, we only update the values of a node's `input_qspec_map`, not the keys. As a result, the keys still refer to the old nodes that do not exist anymore after the QAT fusion. **Test Plan:** ``` python test/quantization/pt2e/test_quantize_pt2e_qat.py -k test_qat_shared_qspec ```
d76110c to
c19b8cd
Compare
namgyu-youn
pushed a commit
to namgyu-youn/ao
that referenced
this pull request
Nov 21, 2025
**Summary:** In pt2e QAT, we first annotate the nodes to be quantized and then perform the pattern replacement (i.e. QAT fusion) in the prepare step. After this pattern replacement, old nodes are replaced with new nodes, and any references to the old nodes must also be updated to refer to the new nodes. This commit fixes a bug where, for special qspecs like the `SharedQuantizationSpec` and `DerivedQuantizationSpec`, we only update the values of a node's `input_qspec_map`, not the keys. As a result, the keys still refer to the old nodes that do not exist anymore after the QAT fusion. **Test Plan:** ``` python test/quantization/pt2e/test_quantize_pt2e_qat.py -k test_qat_shared_qspec ```
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.
Summary: In pt2e QAT, we first annotate the nodes to be quantized and then perform the pattern replacement (i.e. QAT fusion) in the prepare step. After this pattern replacement, old nodes are replaced with new nodes, and any references to the old nodes must also be updated to refer to the new nodes.
This commit fixes a bug where, for special qspecs like the
SharedQuantizationSpecandDerivedQuantizationSpec, we only update the values of a node'sinput_qspec_map, not the keys. As a result, the keys still refer to the old nodes that do not exist anymore after the QAT fusion.Test Plan: