Test_Caffe_nets.FasterRCNN_vgg16 fails with ENABLE_FAST_MATH option #24016
Description
System Information
OpenCV: current 4.x (4.8.0-dev)
Platform: Ubuntu 18.04, GCC 7.5
Hardware: Core i5 2500k (no avx2 and no avx512)
Detailed description
Net inference accuracy issue is raised.
Steps to reproduce
cmake -DENABLE_FAST_MATH=1 ../opencv
make -j4
./bin/opencv_test_dnn --gtest_filter="Test_Caffe_nets.FasterRCNN_vgg16/*"
Log:
[ RUN ] Test_Caffe_nets.FasterRCNN_vgg16/2, where GetParam() = OCV/CPU
Unmatched prediction: class 7 score 0.997029 box [235.421 x 83.0574 from (483.244, 92.7312)]
Highest IoU: 0.970777
/home/alexander/Projects/OpenCV/opencv-master/modules/dnn/test/test_common.impl.hpp:145: Failure
Value of: matched
Actual: false
Expected: true
model name: VGG16_faster_rcnn_final.caffemodel
Unmatched prediction: class 12 score 0.993107 box [213.524 x 372.432 from (136.32, 189.653)]
Highest IoU: 0
/home/alexander/Projects/OpenCV/opencv-master/modules/dnn/test/test_common.impl.hpp:145: Failure
Value of: matched
Actual: false
Expected: true
model name: VGG16_faster_rcnn_final.caffemodel
Unmatched prediction: class 2 score 0.948399 box [502.178 x 251.37 from (98.8219, 211.123)]
Highest IoU: 0
/home/alexander/Projects/OpenCV/opencv-master/modules/dnn/test/test_common.impl.hpp:145: Failure
Value of: matched
Actual: false
Expected: true
model name: VGG16_faster_rcnn_final.caffemodel
Unmatched reference: class 2 score 0.949398 box [501.96 x 252.708 from (99.2454, 210.141)] IoU diff: 1
/home/alexander/Projects/OpenCV/opencv-master/modules/dnn/test/test_common.impl.hpp:157: Failure
Expected: (refScores[i]) <= (confThreshold), actual: 0.949398 vs 0.8
model name: VGG16_faster_rcnn_final.caffemodel
Unmatched reference: class 7 score 0.997022 box [240.844 x 83.6312 from (481.841, 92.3218)] IoU diff: 0.0292231
/home/alexander/Projects/OpenCV/opencv-master/modules/dnn/test/test_common.impl.hpp:157: Failure
Expected: (refScores[i]) <= (confThreshold), actual: 0.997022 vs 0.8
model name: VGG16_faster_rcnn_final.caffemodel
Unmatched reference: class 12 score 0.993028 box [217.773 x 373.789 from (133.221, 189.377)] IoU diff: 1
/home/alexander/Projects/OpenCV/opencv-master/modules/dnn/test/test_common.impl.hpp:157: Failure
Expected: (refScores[i]) <= (confThreshold), actual: 0.993028 vs 0.8
model name: VGG16_faster_rcnn_final.caffemodel
[ FAILED ] Test_Caffe_nets.FasterRCNN_vgg16/2, where GetParam() = OCV/CPU (9083 ms)
Issue submission checklist
- I report the issue, it's not a question
- I checked the problem with documentation, FAQ, open issues, forum.opencv.org, Stack Overflow, etc and have not found any solution
- I updated to the latest OpenCV version and the issue is still there
- There is reproducer code and related data files (videos, images, onnx, etc)