[quant][graphmode] quant_fusion jit pass #24427
Closed
Conversation
This file contains 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
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
This was referenced Aug 15, 2019
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
This was referenced Aug 22, 2019
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
|
What is the order? First we do fusion and then quant-dequant or otherway. |
|
I think in future you will have more fusion patterns. How will the QuantFusion look like ? Should you not iterate over set of patterns ? I will let Mikhail provide more comments as he is better familiar with the fusion patterns |
conv->relu fusion will happen here as well, this PR is just used as a placeholder for quant_fusion pass, and we'll extend this later. |
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
ZolotukhinM
reviewed
Aug 23, 2019
There was a problem hiding this comment.
@nishantpdce, when we have several patterns, they will be matched in order they are registered, so it's up to us to decide what we want first. In our case I'd expect the order will be from bigger to smaller, e.g:
- DQ-Conv-Relu-Q --> QConvRelu
- Conv-Relu --> ConvRelu
IOW, we would like to first check patterns that include Q/DQ before we check patterns from generic fusion.
ZolotukhinM
approved these changes
Aug 23, 2019
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Closed
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags:
|
This pull request has been merged in 9a9ef21. |
xxtEchjovs44
pushed a commit
to xxtEchjovs44/pytorch
that referenced
this pull request
Jan 29, 2020
Summary: Added following pass: - _jit_pass_quant_fusion: Fusion pass that replaces the dequant->conv->quant patterns to quantized_conv Test Plan: python test/test_jit.py Reviewers: pt1quant Subscribers: Tasks: Tags: ghstack-source-id: 34aa4a870d51a63137b37f39773c0081b1b4b99d Pull Request resolved: pytorch/pytorch#24427
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.
Stack from ghstack:
Summary:
Added following pass:
to quantized_conv
Test Plan:
python test/test_jit.py
Reviewers:
pt1quant
Subscribers:
Tasks:
Tags:
Differential Revision: D17001142