Support ONNX operator QLinearSoftmax in dnn#23655
Merged
asmorkalov merged 8 commits intoopencv:4.xfrom May 25, 2023
Merged
Conversation
asmorkalov
reviewed
May 22, 2023
dkurt
reviewed
May 23, 2023
dkurt
reviewed
May 23, 2023
| TEST_P(Test_ONNX_layers, QLinearSoftmax) | ||
| { | ||
| // threshold is set for fusion with dequantization | ||
| testONNXModels("qlinearsoftmax_11", npy, 0.001, 0.002); |
Member
There was a problem hiding this comment.
Does this test cover all 4 scenarios?
SoftmaxInt8Invoker<true>::run(src, dst, blobs[0], N, D, output_sc, output_zp);
SoftmaxInt8Invoker<false>::run(src, dst, blobs[0], N, D, output_sc, output_zp);
SoftmaxInt8OutputFloatInvoker<true>::run(src, dst, blobs[0], N, D);
SoftmaxInt8OutputFloatInvoker<false>::run(src, dst, blobs[0], N, D);
Member
Author
There was a problem hiding this comment.
This test does not cover log softmax, but it is tested by the existing test Test_Int8_layers.Softmax_log_ONNX/0 instead.
The model architecture of this test is like [input]->QuantizeLinear -> QLinearSoftmax -> DequantizeLinear -> [output]. Since we have tryFuse in SoftmaxInt8 which fuses DequantizeLinear and therefore we have output in float instead of int8 in the end, SoftmaxInt8OutputFloatInvoker<false>::run(...) is triggered here in the test. I tested locally with tryFuse disabled which triggers SoftmaxInt8Invoker<false>::run(...) and it passes. Do you think it necessary to add another test case for int8 output like [input] -> QLinearSoftmax -> [output]?
Member
Author
|
All todo items are clear. Please review if possible. |
zihaomu
reviewed
May 25, 2023
zihaomu
approved these changes
May 25, 2023
dkurt
approved these changes
May 25, 2023
Merged
thewoz
pushed a commit
to thewoz/opencv
that referenced
this pull request
Jan 4, 2024
Support ONNX operator QLinearSoftmax in dnn opencv#23655 Resolves opencv#23636. Merge with opencv/opencv_extra#1064. This PR maps the QLinearSoftmax (from com.microsoft domain) to SoftmaxInt8 in dnn along with some speed optimization. Todo: - [x] support QLinearSoftmax with opset = 13 - [x] add model and test data for QLinearSoftmax with opset = 13 - [x] ensure all models have dims >= 3. - [x] add the script to generate model and test data ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
48 tasks
thewoz
pushed a commit
to thewoz/opencv
that referenced
this pull request
May 29, 2024
Support ONNX operator QLinearSoftmax in dnn opencv#23655 Resolves opencv#23636. Merge with opencv/opencv_extra#1064. This PR maps the QLinearSoftmax (from com.microsoft domain) to SoftmaxInt8 in dnn along with some speed optimization. Todo: - [x] support QLinearSoftmax with opset = 13 - [x] add model and test data for QLinearSoftmax with opset = 13 - [x] ensure all models have dims >= 3. - [x] add the script to generate model and test data ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [x] I agree to contribute to the project under Apache 2 License. - [x] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [x] The PR is proposed to the proper branch - [x] There is a reference to the original bug report and related work - [x] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [x] The feature is well documented and sample code can be built with the project CMake
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
Resolves #23636.
Merge with opencv/opencv_extra#1064.
This PR maps the QLinearSoftmax (from com.microsoft domain) to SoftmaxInt8 in dnn along with some speed optimization.
Todo:
Pull Request Readiness Checklist
See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request
Patch to opencv_extra has the same branch name.