Add a pass to insert QDQ nodes before residual connection #59009
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
yghstill
previously approved these changes
Nov 23, 2023
|
This PR only supports distributed optimizer to insert QDQ node before skip-connection, but this should be implemented in QAT API (ex. quant_aware or quantization_pass). I will provide a complete version for this. Marked as draft. |
wanghaoshuang
approved these changes
Dec 4, 2023
SigureMo
pushed a commit
to gouzil/Paddle
that referenced
this pull request
Dec 5, 2023
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Outdated suggestions cannot be applied.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.
PR types
Bug fixes
PR changes
Others
Description
This pull request adds a new pass
AddQuantDequantForResidualinquantization_pass.py.Through this pass,
quant_awarecould insert QDQ nodes for residual connections to ensure that INT8 inference runs entirely under low precision. Otherwise, some kernels may have floating-point precision and intermediate tensors.This PR is an example for issue The model quantized by QAT API should have QDQ nodes before skip connection.