Integrate Automated QDQ placement tool - Part 4 #704
+2,354
−1
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
willg-nv
changed the title
This was referenced Dec 17, 2025
Author
|
@ChenhanYu @cjluo-nv could you help me review this PR? thanks! |
willg-nv
force-pushed
the
dev-willg-integrate-auto-qdq-placement-part4
branch
from
1698082 to
6d55fcb
Compare
gcunhase
reviewed
Jan 9, 2026
|
|
||
| **Q: Can I optimize for accuracy instead of latency?** | ||
|
|
||
| A: Currently, the autotuner optimizes for latency. For accuracy-aware optimization, you would need to implement a custom benchmarking function that evaluates accuracy on a validation dataset. |
Contributor
There was a problem hiding this comment.
Can we provide an example on how users could do that? You may re-use or modify the evaluate.py script in modelopt/examples/onnx_ptq as a starting point.
Author
There was a problem hiding this comment.
Sorry, this line was generated by cursor. This tool currently only focus on perf, accuracy-aware optimization is not supported.
Contributor
There was a problem hiding this comment.
Is there any way for users to implement something to direct the Q/DQ node placement according to an accuracy metric as well or would that not be straight-forward to do?
Signed-off-by: Will Guo <[email protected]>
willg-nv
force-pushed
the
dev-willg-integrate-auto-qdq-placement-part4
branch
from
6d55fcb to
14ab5b6
Compare
gcunhase
reviewed
Jan 13, 2026
|
|
||
| **Q: Can I optimize for accuracy instead of latency?** | ||
|
|
||
| A: Currently, the autotuner optimizes for latency. |
Contributor
There was a problem hiding this comment.
nit: A: Currently, the autotuner optimizes for latency only.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
What does this PR do?
Type of change: new feature
Overview: This PR integrates automated QDQ placement tool to ModelOpt, this PR is 4/4 parts of the changes. This PR contains the following changes:
Part 1: #701
Part 2: #702
Part 3: #703
Part 4: #704
Usage
Testing
This PR does not contains tests
Before your PR is "Ready for review"
Additional Information