mirror/opencv
1
0
Fork 0
mirror of https://github.com/opencv/opencv.git synced 2024-12-15 18:09:11 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity
81f33103fd
opencv/modules/dnn/test/test_common.hpp
Aliaksei Urbanski 35ca2f78d6
Merge pull request #25880 from Jamim:fix/cuda-no-fp16 
Fix CUDA for old GPUs without FP16 support #25880

Fixes #21461

~This is a build-time solution that reflects https://github.com/opencv/opencv/blob/4.10.0/modules/dnn/src/cuda4dnn/init.hpp#L68-L82.~
~We shouldn't add an invalid target while building with `CUDA_ARCH_BIN` < 53.~
_(please see [this discussion](https://github.com/opencv/opencv/pull/25880#discussion_r1668074505))_

This is a run-time solution that basically reverts [these lines](d0fe6ad109 (diff-757c5ab6ddf2f99cdd09f851e3cf17abff203aff4107d908c7ad3d0466f39604L245-R245)).

I've debugged these changes, [coupled with other fixes](https://github.com/gentoo/gentoo/pull/37479), on [Gentoo Linux](https://www.gentoo.org/) and [related tests passed](https://github.com/user-attachments/files/16135391/opencv-4.10.0.20240708-224733.log.gz) on my laptop with `GeForce GTX 960M`.

Alternative solution:
  - #21462

_Best regards!_

### Pull Request Readiness Checklist

- [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
- [ ] `n/a` There is accuracy test, performance test and test data in opencv_extra repository, if applicable
- [ ] `n/a` The feature is well documented and sample code can be built with the project CMake
2024-07-10 12:39:30 +03:00

248 lines
9.3 KiB
C++
Raw Blame History

// This file is part of OpenCV project.
// It is subject to the license terms in the LICENSE file found in the top-level directory
// of this distribution and at http://opencv.org/license.html.
#ifndef __OPENCV_TEST_COMMON_HPP__
#define __OPENCV_TEST_COMMON_HPP__
#include "opencv2/dnn/utils/inference_engine.hpp"
#ifdef HAVE_OPENCL
#include "opencv2/core/ocl.hpp"
#endif
// src/op_inf_engine.hpp
#define INF_ENGINE_VER_MAJOR_GT(ver) (((INF_ENGINE_RELEASE) / 10000) > ((ver) / 10000))
#define INF_ENGINE_VER_MAJOR_GE(ver) (((INF_ENGINE_RELEASE) / 10000) >= ((ver) / 10000))
#define INF_ENGINE_VER_MAJOR_LT(ver) (((INF_ENGINE_RELEASE) / 10000) < ((ver) / 10000))
#define INF_ENGINE_VER_MAJOR_LE(ver) (((INF_ENGINE_RELEASE) / 10000) <= ((ver) / 10000))
#define INF_ENGINE_VER_MAJOR_EQ(ver) (((INF_ENGINE_RELEASE) / 10000) == ((ver) / 10000))
#define CV_TEST_TAG_DNN_SKIP_OPENCV_BACKEND "dnn_skip_opencv_backend"
#define CV_TEST_TAG_DNN_SKIP_HALIDE "dnn_skip_halide"
#define CV_TEST_TAG_DNN_SKIP_CPU "dnn_skip_cpu"
#define CV_TEST_TAG_DNN_SKIP_CPU_FP16 "dnn_skip_cpu_fp16"
#define CV_TEST_TAG_DNN_SKIP_OPENCL "dnn_skip_ocl"
#define CV_TEST_TAG_DNN_SKIP_OPENCL_FP16 "dnn_skip_ocl_fp16"
#define CV_TEST_TAG_DNN_SKIP_IE_NN_BUILDER "dnn_skip_ie_nn_builder"
#define CV_TEST_TAG_DNN_SKIP_IE_NGRAPH "dnn_skip_ie_ngraph"
#define CV_TEST_TAG_DNN_SKIP_IE "dnn_skip_ie"
#define CV_TEST_TAG_DNN_SKIP_IE_2018R5 "dnn_skip_ie_2018r5"
#define CV_TEST_TAG_DNN_SKIP_IE_2019R1 "dnn_skip_ie_2019r1"
#define CV_TEST_TAG_DNN_SKIP_IE_2019R1_1 "dnn_skip_ie_2019r1_1"
#define CV_TEST_TAG_DNN_SKIP_IE_2019R2 "dnn_skip_ie_2019r2"
#define CV_TEST_TAG_DNN_SKIP_IE_2019R3 "dnn_skip_ie_2019r3"
#define CV_TEST_TAG_DNN_SKIP_IE_CPU "dnn_skip_ie_cpu"
#define CV_TEST_TAG_DNN_SKIP_IE_OPENCL "dnn_skip_ie_ocl"
#define CV_TEST_TAG_DNN_SKIP_IE_OPENCL_FP16 "dnn_skip_ie_ocl_fp16"
#define CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_2 "dnn_skip_ie_myriad2"
#define CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_X "dnn_skip_ie_myriadx"
#define CV_TEST_TAG_DNN_SKIP_IE_MYRIAD CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_2, CV_TEST_TAG_DNN_SKIP_IE_MYRIAD_X
#define CV_TEST_TAG_DNN_SKIP_IE_ARM_CPU "dnn_skip_ie_arm_cpu"
#define CV_TEST_TAG_DNN_SKIP_VULKAN "dnn_skip_vulkan"
#define CV_TEST_TAG_DNN_SKIP_CUDA "dnn_skip_cuda"
#define CV_TEST_TAG_DNN_SKIP_CUDA_FP16 "dnn_skip_cuda_fp16"
#define CV_TEST_TAG_DNN_SKIP_CUDA_FP32 "dnn_skip_cuda_fp32"
#define CV_TEST_TAG_DNN_SKIP_ONNX_CONFORMANCE "dnn_skip_onnx_conformance"
#define CV_TEST_TAG_DNN_SKIP_PARSER "dnn_skip_parser"
#define CV_TEST_TAG_DNN_SKIP_GLOBAL "dnn_skip_global"
#define CV_TEST_TAG_DNN_SKIP_TIMVX "dnn_skip_timvx"
#define CV_TEST_TAG_DNN_SKIP_CANN "dnn_skip_cann"
#ifdef HAVE_INF_ENGINE
#if INF_ENGINE_VER_MAJOR_EQ(2018050000)
# define CV_TEST_TAG_DNN_SKIP_IE_VERSION CV_TEST_TAG_DNN_SKIP_IE, CV_TEST_TAG_DNN_SKIP_IE_2018R5
#elif INF_ENGINE_VER_MAJOR_EQ(2019010000)
# if INF_ENGINE_RELEASE < 2019010100
# define CV_TEST_TAG_DNN_SKIP_IE_VERSION CV_TEST_TAG_DNN_SKIP_IE, CV_TEST_TAG_DNN_SKIP_IE_2019R1
# else
# define CV_TEST_TAG_DNN_SKIP_IE_VERSION CV_TEST_TAG_DNN_SKIP_IE, CV_TEST_TAG_DNN_SKIP_IE_2019R1_1
# endif
#elif INF_ENGINE_VER_MAJOR_EQ(2019020000)
# define CV_TEST_TAG_DNN_SKIP_IE_VERSION CV_TEST_TAG_DNN_SKIP_IE, CV_TEST_TAG_DNN_SKIP_IE_2019R2
#elif INF_ENGINE_VER_MAJOR_EQ(2019030000)
# define CV_TEST_TAG_DNN_SKIP_IE_VERSION CV_TEST_TAG_DNN_SKIP_IE, CV_TEST_TAG_DNN_SKIP_IE_2019R3
#endif
#endif // HAVE_INF_ENGINE
#ifndef CV_TEST_TAG_DNN_SKIP_IE_VERSION
# define CV_TEST_TAG_DNN_SKIP_IE_VERSION CV_TEST_TAG_DNN_SKIP_IE
#endif
namespace cv { namespace dnn {
CV__DNN_INLINE_NS_BEGIN
void PrintTo(const cv::dnn::Backend& v, std::ostream* os);
void PrintTo(const cv::dnn::Target& v, std::ostream* os);
using opencv_test::tuple;
using opencv_test::get;
void PrintTo(const tuple<cv::dnn::Backend, cv::dnn::Target> v, std::ostream* os);
CV__DNN_INLINE_NS_END
}} // namespace cv::dnn
namespace opencv_test {
void initDNNTests();
using namespace cv::dnn;
static inline const std::string &getOpenCVExtraDir()
{
return cvtest::TS::ptr()->get_data_path();
}
void normAssert(
cv::InputArray ref, cv::InputArray test, const char *comment = "",
double l1 = 0.00001, double lInf = 0.0001);
std::vector<cv::Rect2d> matToBoxes(const cv::Mat& m);
void normAssertDetections(
const std::vector<int>& refClassIds,
const std::vector<float>& refScores,
const std::vector<cv::Rect2d>& refBoxes,
const std::vector<int>& testClassIds,
const std::vector<float>& testScores,
const std::vector<cv::Rect2d>& testBoxes,
const char *comment = "", double confThreshold = 0.0,
double scores_diff = 1e-5, double boxes_iou_diff = 1e-4);
// For SSD-based object detection networks which produce output of shape 1x1xNx7
// where N is a number of detections and an every detection is represented by
// a vector [batchId, classId, confidence, left, top, right, bottom].
void normAssertDetections(
cv::Mat ref, cv::Mat out, const char *comment = "",
double confThreshold = 0.0, double scores_diff = 1e-5,
double boxes_iou_diff = 1e-4);
// For text detection networks
// Curved text polygon is not supported in the current version.
// (concave polygon is invalid input to intersectConvexConvex)
void normAssertTextDetections(
const std::vector<std::vector<Point>>& gtPolys,
const std::vector<std::vector<Point>>& testPolys,
const char *comment = "", double boxes_iou_diff = 1e-4);
void readFileContent(const std::string& filename, CV_OUT std::vector<char>& content);
bool validateVPUType();
testing::internal::ParamGenerator< tuple<Backend, Target> > dnnBackendsAndTargets(
bool withInferenceEngine = true,
bool withHalide = false,
bool withCpuOCV = true,
bool withVkCom = true,
bool withCUDA = true,
bool withNgraph = true,
bool withWebnn = true,
bool withCann = true
);
testing::internal::ParamGenerator< tuple<Backend, Target> > dnnBackendsAndTargetsIE();
class DNNTestLayer : public TestWithParam<tuple<Backend, Target> >
{
public:
dnn::Backend backend;
dnn::Target target;
double default_l1, default_lInf;
DNNTestLayer()
{
backend = (dnn::Backend)(int)get<0>(GetParam());
target = (dnn::Target)(int)get<1>(GetParam());
getDefaultThresholds(backend, target, &default_l1, &default_lInf);
}
static void getDefaultThresholds(int backend, int target, double* l1, double* lInf)
{
if (target == DNN_TARGET_CPU_FP16 || target == DNN_TARGET_CUDA_FP16 || target == DNN_TARGET_OPENCL_FP16 || target == DNN_TARGET_MYRIAD)
{
*l1 = 4e-3;
*lInf = 2e-2;
}
else
{
*l1 = 1e-5;
*lInf = 1e-4;
}
}
static void checkBackend(int backend, int target, Mat* inp = 0, Mat* ref = 0)
{
CV_UNUSED(backend); CV_UNUSED(target); CV_UNUSED(inp); CV_UNUSED(ref);
#if defined(INF_ENGINE_RELEASE) && INF_ENGINE_VER_MAJOR_LT(2021000000)
if ((backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019 || backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH)
&& target == DNN_TARGET_MYRIAD)
{
if (inp && ref && inp->dims == 4 && ref->dims == 4 &&
inp->size[0] != 1 && inp->size[0] != ref->size[0])
{
std::cout << "Inconsistent batch size of input and output blobs for Myriad plugin" << std::endl;
applyTestTag(CV_TEST_TAG_DNN_SKIP_IE_MYRIAD);
}
}
#endif
}
void expectNoFallbacks(Net& net, bool raiseError = true)
{
// Check if all the layers are supported with current backend and target.
// Some layers might be fused so their timings equal to zero.
std::vector<double> timings;
net.getPerfProfile(timings);
std::vector<String> names = net.getLayerNames();
CV_Assert(names.size() == timings.size());
bool hasFallbacks = false;
for (int i = 0; i < names.size(); ++i)
{
Ptr<dnn::Layer> l = net.getLayer(net.getLayerId(names[i]));
bool fused = !timings[i];
if ((!l->supportBackend(backend) || l->preferableTarget != target) && !fused)
{
hasFallbacks = true;
std::cout << "FALLBACK: Layer [" << l->type << "]:[" << l->name << "] is expected to have backend implementation" << endl;
}
}
if (hasFallbacks && raiseError)
CV_Error(Error::StsNotImplemented, "Implementation fallbacks are not expected in this test");
}
void expectNoFallbacksFromIE(Net& net)
{
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NN_BUILDER_2019)
expectNoFallbacks(net);
if (backend == DNN_BACKEND_INFERENCE_ENGINE_NGRAPH)
expectNoFallbacks(net, false);
}
void expectNoFallbacksFromCUDA(Net& net)
{
if (backend == DNN_BACKEND_CUDA)
expectNoFallbacks(net);
}
size_t getTopMemoryUsageMB();
protected:
void checkBackend(Mat* inp = 0, Mat* ref = 0)
{
checkBackend(backend, target, inp, ref);
}
};
} // namespace
#endif
Reference in New Issue View Git Blame Copy Permalink