mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-06-21 02:18:09 +00:00
Move NMS to Onnx domain (#865)
* move files * move files * Remove NonMaxSuppression from Contrib op, move it to Onnx domain, opset 10 * move NMS out of namespace contrib * update data type in UT * update to latest onnx * white list the node test for Mod which is not implemented yet
This commit is contained in:
Hector Li
GitHub
parent
2947e1f9d4
commit
e8d722003a
11 changed files with 83 additions and 137 deletions
2
cmake/external/onnx
vendored
2
cmake/external/onnx
vendored
|
|
@ -1 +1 @@
|
|||
Subproject commit 83dd62659fc07d5b7fa93b5d1c1879f93509c7db
|
||||
Subproject commit 0e8d2bc5e51455c70ef790b9f65aa632ed9bc8a7
|
||||
|
|
@ -13,7 +13,6 @@ class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1,
|
|||
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, float, FusedGemm);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, AttnLSTM);
|
||||
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, string, Tokenizer);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, NonMaxSuppression);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, Range);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, WordConvEmbedding);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, GatherND);
|
||||
|
|
@ -59,7 +58,6 @@ void RegisterContribKernels(KernelRegistry& kernel_registry) {
|
|||
BuildKernelCreateInfo<ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, float, FusedGemm)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, AttnLSTM)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, string, Tokenizer)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, NonMaxSuppression)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, Range)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, WordConvEmbedding)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kMSDomain, 1, GatherND)>,
|
||||
|
|
|
|||
|
|
@ -875,64 +875,6 @@ with the exception that numpy default keepdims to False instead of True.)DOC")
|
|||
"Keep the reduced dimension or not, default 1 mean keep reduced dimension.",
|
||||
AttributeProto::INT);
|
||||
|
||||
ONNX_CONTRIB_OPERATOR_SCHEMA(NonMaxSuppression)
|
||||
.SetDomain(kMSDomain)
|
||||
.SinceVersion(1)
|
||||
.SetDoc(R"DOC(
|
||||
Filter out boxes that have high intersection-over-union (IOU) overlap with previously selected boxes.
|
||||
Bounding boxes with score less than score_threshold are removed. Bounding box format is indicated by attribute center_point_box.
|
||||
Note that this algorithm is agnostic to where the origin is in the coordinate system and more generally is invariant to
|
||||
orthogonal transformations and translations of the coordinate system; thus translating or reflections of the coordinate system
|
||||
result in the same boxes being selected by the algorithm.
|
||||
The selected_indices output is a set of integers indexing into the input collection of bounding boxes representing the selected boxes.
|
||||
The bounding box coordinates corresponding to the selected indices can then be obtained using the Gather or GatherND operation.
|
||||
Note: The boxes doesn't has class dimension which means it alwasy has scores calculated for different classes on same box.)DOC")
|
||||
.Input(
|
||||
0,
|
||||
"boxes",
|
||||
"An input tensor with shape [num_batches, spatial_dimension, 4]. The single box data format is indicated by center_point_box.",
|
||||
"tensor(float)")
|
||||
.Input(
|
||||
1,
|
||||
"scores",
|
||||
"An input tensor with shape [num_batches, num_classes, spatial_dimension]",
|
||||
"tensor(float)")
|
||||
.Input(
|
||||
2,
|
||||
"max_output_boxes_per_class",
|
||||
"Integer representing the maximum number of boxes to be selected per batch per class. It is a scalar.",
|
||||
"tensor(int32)",
|
||||
OpSchema::Optional)
|
||||
.Input(
|
||||
3,
|
||||
"iou_threshold",
|
||||
"Float representing the threshold for deciding whether boxes overlap too much with respect to IOU. It is scalar. Value range [0, 1].",
|
||||
"tensor(float)",
|
||||
OpSchema::Optional)
|
||||
.Input(
|
||||
4,
|
||||
"score_threshold",
|
||||
"Float representing the threshold for deciding when to remove boxes based on score. It is a scalar",
|
||||
"tensor(float)",
|
||||
OpSchema::Optional)
|
||||
.Output(
|
||||
0,
|
||||
"selected_indices",
|
||||
"selected indices from the boxes tensor. [num_selected_indices, 3], the selected indices format is [batch_index, class_index, box_index].",
|
||||
"tensor(int32)")
|
||||
.Attr(
|
||||
"center_point_box",
|
||||
"Integer indicate the format of the box data. The default is 0."
|
||||
"0 - the box data is supplied as [y1, x1, y2, x2] where (y1, x1) and (y2, x2) are the coordinates of any diagonal pair of box corners"
|
||||
"and the coordinates can be provided as normalized (i.e., lying in the interval [0, 1]) or absolute. Mostly used for TF models."
|
||||
"1 - the box data is supplied as [x_center, y_center, width, height]. Mostly used for Pytoch models.",
|
||||
AttributeProto::INT,
|
||||
static_cast<int64_t>(0))
|
||||
.TypeAndShapeInferenceFunction([](ONNX_NAMESPACE::InferenceContext& ctx) {
|
||||
auto selected_indices_type = ctx.getOutputType(0)->mutable_tensor_type();
|
||||
selected_indices_type->set_elem_type(::ONNX_NAMESPACE::TensorProto_DataType::TensorProto_DataType_INT32);
|
||||
});
|
||||
|
||||
ONNX_CONTRIB_OPERATOR_SCHEMA(MurmurHash3)
|
||||
.SetDomain(kMSDomain)
|
||||
.SinceVersion(1)
|
||||
|
|
|
|||
|
|
@ -264,6 +264,7 @@ class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain,
|
|||
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, int64_t, Slice);
|
||||
class ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, string, Slice);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, Dropout);
|
||||
class ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, NonMaxSuppression);
|
||||
|
||||
void RegisterOnnxOperatorKernels(KernelRegistry& kernel_registry) {
|
||||
static const BuildKernelCreateInfoFn function_table[] = {
|
||||
|
|
@ -517,6 +518,7 @@ void RegisterOnnxOperatorKernels(KernelRegistry& kernel_registry) {
|
|||
BuildKernelCreateInfo<ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, int64_t, Slice)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_TYPED_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, string, Slice)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, Dropout)>,
|
||||
BuildKernelCreateInfo<ONNX_OPERATOR_KERNEL_CLASS_NAME(kCpuExecutionProvider, kOnnxDomain, 10, NonMaxSuppression)>,
|
||||
};
|
||||
|
||||
for (auto& function_table_entry : function_table) {
|
||||
|
|
|
|||
|
|
@ -11,16 +11,15 @@ limitations under the License.
|
|||
==============================================================================*/
|
||||
/* Modifications Copyright (c) Microsoft. */
|
||||
|
||||
#include "contrib_ops/cpu/non_max_suppression.h"
|
||||
#include "non_max_suppression.h"
|
||||
#include <queue>
|
||||
|
||||
namespace onnxruntime {
|
||||
namespace contrib {
|
||||
|
||||
ONNX_OPERATOR_KERNEL_EX(
|
||||
NonMaxSuppression,
|
||||
kMSDomain,
|
||||
1,
|
||||
kOnnxDomain,
|
||||
10,
|
||||
kCpuExecutionProvider,
|
||||
KernelDefBuilder(),
|
||||
NonMaxSuppression);
|
||||
|
|
@ -35,7 +34,7 @@ void NonMaxSuppression::MaxMin(const float& lhs, const float& rhs, float& min, f
|
|||
}
|
||||
}
|
||||
|
||||
bool NonMaxSuppression::SuppressByIOU(const float* boxes_data, int32_t box_index1, int32_t box_index2, float iou_threshold) const {
|
||||
bool NonMaxSuppression::SuppressByIOU(const float* boxes_data, int64_t box_index1, int64_t box_index2, float iou_threshold) const {
|
||||
float x1_min, y1_min, x1_max, y1_max, x2_min, y2_min, x2_max, y2_max;
|
||||
// center_point_box_ only support 0 or 1
|
||||
if (0 == center_point_box_) {
|
||||
|
|
@ -88,7 +87,7 @@ bool NonMaxSuppression::SuppressByIOU(const float* boxes_data, int32_t box_index
|
|||
}
|
||||
|
||||
Status NonMaxSuppression::ParepareCompute(OpKernelContext* ctx, const TensorShape& boxes_shape, const TensorShape& scores_shape,
|
||||
int32_t& max_output_boxes_per_class, float& iou_threshold, float& score_threshold, bool& has_score_threshold) const {
|
||||
int64_t& max_output_boxes_per_class, float& iou_threshold, float& score_threshold, bool& has_score_threshold) const {
|
||||
ORT_RETURN_IF_NOT(boxes_shape.NumDimensions() == 3, "boxes must be a 3D tensor.");
|
||||
ORT_RETURN_IF_NOT(scores_shape.NumDimensions() == 3, "scores must be a 3D tensor.");
|
||||
|
||||
|
|
@ -104,7 +103,7 @@ Status NonMaxSuppression::ParepareCompute(OpKernelContext* ctx, const TensorShap
|
|||
|
||||
const Tensor* max_output_boxes_per_class_tensor = ctx->Input<Tensor>(2);
|
||||
if (max_output_boxes_per_class_tensor != nullptr) {
|
||||
max_output_boxes_per_class = *(max_output_boxes_per_class_tensor->Data<int32_t>());
|
||||
max_output_boxes_per_class = *(max_output_boxes_per_class_tensor->Data<int64_t>());
|
||||
max_output_boxes_per_class = max_output_boxes_per_class > 0 ? max_output_boxes_per_class : 0;
|
||||
}
|
||||
|
||||
|
|
@ -132,7 +131,7 @@ Status NonMaxSuppression::Compute(OpKernelContext* ctx) const {
|
|||
auto& boxes_shape = boxes->Shape();
|
||||
auto& scores_shape = scores->Shape();
|
||||
|
||||
int32_t max_output_boxes_per_class = 0;
|
||||
int64_t max_output_boxes_per_class = 0;
|
||||
float iou_threshold = 0;
|
||||
// Not so sure for the value range of score_threshold, so set a bool to indicate whether it has this input
|
||||
bool has_score_threshold = false;
|
||||
|
|
@ -152,7 +151,7 @@ Status NonMaxSuppression::Compute(OpKernelContext* ctx) const {
|
|||
|
||||
struct ScoreIndexPair {
|
||||
float score;
|
||||
int32_t index;
|
||||
int64_t index;
|
||||
};
|
||||
|
||||
auto LessCompare = [](const ScoreIndexPair& lhs, const ScoreIndexPair& rhs) {
|
||||
|
|
@ -168,12 +167,12 @@ Status NonMaxSuppression::Compute(OpKernelContext* ctx) const {
|
|||
std::priority_queue<ScoreIndexPair, std::deque<ScoreIndexPair>, decltype(LessCompare)> sorted_scores_with_index(LessCompare);
|
||||
for (int64_t box_index = 0; box_index < num_boxes_; ++box_index) {
|
||||
if (!has_score_threshold || (has_score_threshold && scores_data[box_score_offset + box_index] > score_threshold)) {
|
||||
sorted_scores_with_index.emplace(ScoreIndexPair({scores_data[box_score_offset + box_index], static_cast<int32_t>(box_index)}));
|
||||
sorted_scores_with_index.emplace(ScoreIndexPair({scores_data[box_score_offset + box_index], box_index}));
|
||||
}
|
||||
}
|
||||
|
||||
ScoreIndexPair next_top_score;
|
||||
std::vector<int32_t> selected_indicies_inside_class;
|
||||
std::vector<int64_t> selected_indicies_inside_class;
|
||||
// Get the next box with top score, filter by iou_threshold_
|
||||
while (!sorted_scores_with_index.empty()) {
|
||||
next_top_score = sorted_scores_with_index.top();
|
||||
|
|
@ -189,11 +188,11 @@ Status NonMaxSuppression::Compute(OpKernelContext* ctx) const {
|
|||
}
|
||||
|
||||
if (selected) {
|
||||
if (max_output_boxes_per_class > 0 && selected_indicies_inside_class.size() >= max_output_boxes_per_class) {
|
||||
if (max_output_boxes_per_class > 0 && static_cast<int64_t>(selected_indicies_inside_class.size()) >= max_output_boxes_per_class) {
|
||||
break;
|
||||
}
|
||||
selected_indicies_inside_class.push_back(next_top_score.index);
|
||||
tmp_selected_indices.push_back(selected_index(static_cast<int32_t>(batch_index), static_cast<int32_t>(class_index), next_top_score.index));
|
||||
tmp_selected_indices.push_back(selected_index(batch_index, class_index, next_top_score.index));
|
||||
}
|
||||
} //while
|
||||
} //for class_index
|
||||
|
|
@ -202,10 +201,9 @@ Status NonMaxSuppression::Compute(OpKernelContext* ctx) const {
|
|||
int32_t num_selected = static_cast<int32_t>(tmp_selected_indices.size());
|
||||
Tensor* selected_indices = ctx->Output(0, {num_selected, 3});
|
||||
ORT_ENFORCE(selected_indices);
|
||||
memcpy(selected_indices->MutableData<int32_t>(), tmp_selected_indices.data(), num_selected * sizeof(selected_index));
|
||||
memcpy(selected_indices->MutableData<int64_t>(), tmp_selected_indices.data(), num_selected * sizeof(selected_index));
|
||||
|
||||
return Status::OK();
|
||||
}
|
||||
|
||||
} // namespace contrib
|
||||
} // namespace onnxruntime
|
||||
|
|
@ -7,7 +7,6 @@
|
|||
#include "core/framework/op_kernel.h"
|
||||
|
||||
namespace onnxruntime {
|
||||
namespace contrib {
|
||||
|
||||
class NonMaxSuppression final : public OpKernel {
|
||||
public:
|
||||
|
|
@ -22,10 +21,10 @@ class NonMaxSuppression final : public OpKernel {
|
|||
Status Compute(OpKernelContext* context) const override;
|
||||
|
||||
private:
|
||||
bool SuppressByIOU(const float* boxes_data, int32_t box_index1, int32_t box_index2, float iou_threshold) const;
|
||||
bool SuppressByIOU(const float* boxes_data, int64_t box_index1, int64_t box_index2, float iou_threshold) const;
|
||||
void MaxMin(const float& lhs, const float& rhs, float& min, float& max) const;
|
||||
Status ParepareCompute(OpKernelContext* ctx, const TensorShape& boxes_shape, const TensorShape& scores_shape,
|
||||
int32_t& max_output_boxes_per_batch, float& iou_threshold, float& score_threshold, bool& has_score_threshold) const;
|
||||
int64_t& max_output_boxes_per_batch, float& iou_threshold, float& score_threshold, bool& has_score_threshold) const;
|
||||
|
||||
private:
|
||||
int64_t center_point_box_;
|
||||
|
|
@ -35,12 +34,11 @@ class NonMaxSuppression final : public OpKernel {
|
|||
int64_t num_boxes_;
|
||||
|
||||
struct selected_index {
|
||||
selected_index(int32_t batch_index, int32_t class_index, int32_t box_index)
|
||||
selected_index(int64_t batch_index, int64_t class_index, int64_t box_index)
|
||||
: batch_index_(batch_index), class_index_(class_index), box_index_(box_index) {}
|
||||
int32_t batch_index_ = 0;
|
||||
int32_t class_index_ = 0;
|
||||
int32_t box_index_ = 0;
|
||||
int64_t batch_index_ = 0;
|
||||
int64_t class_index_ = 0;
|
||||
int64_t box_index_ = 0;
|
||||
};
|
||||
};
|
||||
} // namespace contrib
|
||||
} // namespace onnxruntime
|
||||
|
|
@ -350,7 +350,11 @@ int real_main(int argc, char* argv[], OrtEnv** p_env) {
|
|||
{"shrink", "test case is wrong"},
|
||||
{"maxpool_2d_precomputed_strides", "ShapeInferenceError"},
|
||||
{"averagepool_2d_precomputed_strides", "ShapeInferenceError"},
|
||||
{"maxpool_with_argmax_2d_precomputed_strides", "ShapeInferenceError"}
|
||||
{"maxpool_with_argmax_2d_precomputed_strides", "ShapeInferenceError"},
|
||||
{"test_mod_bcast", "not implemented"},
|
||||
{"test_mod_float_mixed_sign_example", "not implemented"},
|
||||
{"test_mod_fmod_mixed_sign_example", "not implemented"},
|
||||
{"test_mod_int64_mixed_sign_example", "not implemented"}
|
||||
};
|
||||
|
||||
#ifdef USE_CUDA
|
||||
|
|
|
|||
|
|
@ -8,7 +8,7 @@ namespace onnxruntime {
|
|||
namespace test {
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, WithIOUThreshold) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -17,10 +17,10 @@ TEST(NonMaxSuppressionOpTest, WithIOUThreshold) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {3, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {3, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
0L, 0L, 5L});
|
||||
|
|
@ -28,7 +28,7 @@ TEST(NonMaxSuppressionOpTest, WithIOUThreshold) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, CenterPointBoxFormat) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.5f, 0.5f, 1.0f, 1.0f,
|
||||
0.5f, 0.6f, 1.0f, 1.0f,
|
||||
|
|
@ -37,10 +37,10 @@ TEST(NonMaxSuppressionOpTest, CenterPointBoxFormat) {
|
|||
0.5f, 10.6f, 1.0f, 1.0f,
|
||||
0.5f, 100.5f, 1.0f, 1.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {3, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {3, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
0L, 0L, 5L});
|
||||
|
|
@ -49,7 +49,7 @@ TEST(NonMaxSuppressionOpTest, CenterPointBoxFormat) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, TwoClasses) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -60,10 +60,10 @@ TEST(NonMaxSuppressionOpTest, TwoClasses) {
|
|||
test.AddInput<float>("scores", {1, 2, 6},
|
||||
{0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f,
|
||||
0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {6L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {6L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {6, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {6, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
0L, 0L, 5L,
|
||||
|
|
@ -74,7 +74,7 @@ TEST(NonMaxSuppressionOpTest, TwoClasses) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, TwoBathes) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {2, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -92,10 +92,10 @@ TEST(NonMaxSuppressionOpTest, TwoBathes) {
|
|||
test.AddInput<float>("scores", {2, 1, 6},
|
||||
{0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f,
|
||||
0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {2L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {2L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {4, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {4, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
1L, 0L, 3L,
|
||||
|
|
@ -104,7 +104,7 @@ TEST(NonMaxSuppressionOpTest, TwoBathes) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, WithScoreThreshold) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -113,17 +113,17 @@ TEST(NonMaxSuppressionOpTest, WithScoreThreshold) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.4f});
|
||||
test.AddOutput<int32_t>("selected_indices", {2, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {2, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, WithoutScoreThreshold) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -132,9 +132,9 @@ TEST(NonMaxSuppressionOpTest, WithoutScoreThreshold) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddOutput<int32_t>("selected_indices", {3, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {3, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
0L, 0L, 5L});
|
||||
|
|
@ -142,7 +142,7 @@ TEST(NonMaxSuppressionOpTest, WithoutScoreThreshold) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, WithScoreThresholdZeroScores) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -151,17 +151,17 @@ TEST(NonMaxSuppressionOpTest, WithScoreThresholdZeroScores) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.1f, 0.0f, 0.0f, 0.3f, 0.2f, -5.0f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {6L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {6L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {-3.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {2, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {2, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, FlippedCoordinates) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{1.0f, 1.0f, 0.0f, 0.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -170,10 +170,10 @@ TEST(NonMaxSuppressionOpTest, FlippedCoordinates) {
|
|||
1.0f, 10.1f, 0.0f, 11.1f,
|
||||
1.0f, 101.0f, 0.0f, 100.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {3, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {3, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
0L, 0L, 5L});
|
||||
|
|
@ -181,7 +181,7 @@ TEST(NonMaxSuppressionOpTest, FlippedCoordinates) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, SelectTwo) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -190,17 +190,17 @@ TEST(NonMaxSuppressionOpTest, SelectTwo) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {2L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {2L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {2, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {2, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, SelectThirty) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -209,10 +209,10 @@ TEST(NonMaxSuppressionOpTest, SelectThirty) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {30L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {30L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {3, 3},
|
||||
test.AddOutput<int64_t>("selected_indices", {3, 3},
|
||||
{0L, 0L, 3L,
|
||||
0L, 0L, 0L,
|
||||
0L, 0L, 5L});
|
||||
|
|
@ -220,19 +220,19 @@ TEST(NonMaxSuppressionOpTest, SelectThirty) {
|
|||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, SelectSingleBox) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 1, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 1}, {0.9f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {1, 3}, {0L, 0L, 0L});
|
||||
test.AddOutput<int64_t>("selected_indices", {1, 3}, {0L, 0L, 0L});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, SelectFromIdenticalBoxes) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 10, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.0f, 1.0f, 1.0f,
|
||||
|
|
@ -246,15 +246,15 @@ TEST(NonMaxSuppressionOpTest, SelectFromIdenticalBoxes) {
|
|||
0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.0f, 1.0f, 1.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 10}, {0.9f, 0.9f, 0.9f, 0.9f, 0.9f, 0.9f, 0.9f, 0.9f, 0.9f, 0.9f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {1, 3}, {0L, 0L, 0L});
|
||||
test.AddOutput<int64_t>("selected_indices", {1, 3}, {0L, 0L, 0L});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, InconsistentBoxAndScoreShapes) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -263,37 +263,37 @@ TEST(NonMaxSuppressionOpTest, InconsistentBoxAndScoreShapes) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 5}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {30L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {30L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {0, 3}, {});
|
||||
test.AddOutput<int64_t>("selected_indices", {0, 3}, {});
|
||||
test.Run(OpTester::ExpectResult::kExpectFailure, "boxes and scores should have same spatial_dimention.");
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, InvalidIOUThreshold) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 1, 4}, {0.0f, 0.0f, 1.0f, 1.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 1}, {0.9f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {3L});
|
||||
test.AddInput<float>("iou_threshold", {}, {1.2f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {0, 3}, {});
|
||||
test.AddOutput<int64_t>("selected_indices", {0, 3}, {});
|
||||
test.Run(OpTester::ExpectResult::kExpectFailure, "iou_threshold must be in range [0, 1]");
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, EmptyInput) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 0, 4}, {});
|
||||
test.AddInput<float>("scores", {1, 1, 0}, {});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {30L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {30L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.0f});
|
||||
test.AddOutput<int32_t>("selected_indices", {0, 3}, {});
|
||||
test.AddOutput<int64_t>("selected_indices", {0, 3}, {});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
TEST(NonMaxSuppressionOpTest, ZeroMaxOutputPerClass) {
|
||||
OpTester test("NonMaxSuppression", 1, onnxruntime::kMSDomain);
|
||||
OpTester test("NonMaxSuppression", 10, kOnnxDomain);
|
||||
test.AddInput<float>("boxes", {1, 6, 4},
|
||||
{0.0f, 0.0f, 1.0f, 1.0f,
|
||||
0.0f, 0.1f, 1.0f, 1.1f,
|
||||
|
|
@ -302,10 +302,10 @@ TEST(NonMaxSuppressionOpTest, ZeroMaxOutputPerClass) {
|
|||
0.0f, 10.1f, 1.0f, 11.1f,
|
||||
0.0f, 100.0f, 1.0f, 101.0f});
|
||||
test.AddInput<float>("scores", {1, 1, 6}, {0.9f, 0.75f, 0.6f, 0.95f, 0.5f, 0.3f});
|
||||
test.AddInput<int32_t>("max_output_boxes_per_class", {}, {0L});
|
||||
test.AddInput<int64_t>("max_output_boxes_per_class", {}, {0L});
|
||||
test.AddInput<float>("iou_threshold", {}, {0.5f});
|
||||
test.AddInput<float>("score_threshold", {}, {0.4f});
|
||||
test.AddOutput<int32_t>("selected_indices", {0, 3}, {});
|
||||
test.AddOutput<int64_t>("selected_indices", {0, 3}, {});
|
||||
test.Run();
|
||||
}
|
||||
|
||||
|
|
@ -95,6 +95,10 @@ backend_test.exclude(r'('
|
|||
'|^test_isinf_cpu.*'
|
||||
'|^test_isinf_negative_cpu.*'
|
||||
'|^test_isinf_positive_cpu.*'
|
||||
'|^test_mod_bcast.*'
|
||||
'|^test_mod_float_mixed_sign_example.*'
|
||||
'|^test_mod_fmod_mixed_sign_example.*'
|
||||
'|^test_mod_int64_mixed_sign_example.*'
|
||||
')')
|
||||
|
||||
# import all test cases at global scope to make
|
||||
|
|
|
|||
|
|
@ -38,8 +38,8 @@ else
|
|||
#5af210ca8a1c73aa6bae8754c9346ec54d0a756e is v1.2.3
|
||||
#bae6333e149a59a3faa9c4d9c44974373dcf5256 is v1.3.0
|
||||
#9e55ace55aad1ada27516038dfbdc66a8a0763db is v1.4.1
|
||||
#83dd62659fc07d5b7fa93b5d1c1879f93509c7db is v1.4.1 latest
|
||||
for onnx_version in "5af210ca8a1c73aa6bae8754c9346ec54d0a756e" "bae6333e149a59a3faa9c4d9c44974373dcf5256" "9e55ace55aad1ada27516038dfbdc66a8a0763db" "83dd62659fc07d5b7fa93b5d1c1879f93509c7db"; do
|
||||
#0e8d2bc5e51455c70ef790b9f65aa632ed9bc8a7 is v1.4.1 latest
|
||||
for onnx_version in "5af210ca8a1c73aa6bae8754c9346ec54d0a756e" "bae6333e149a59a3faa9c4d9c44974373dcf5256" "9e55ace55aad1ada27516038dfbdc66a8a0763db" "0e8d2bc5e51455c70ef790b9f65aa632ed9bc8a7"; do
|
||||
if [ -z ${lastest_onnx_version+x} ]; then
|
||||
echo "first pass";
|
||||
else
|
||||
|
|
|
|||
|
|
@ -32,8 +32,8 @@ else
|
|||
#5af210ca8a1c73aa6bae8754c9346ec54d0a756e is v1.2.3
|
||||
#bae6333e149a59a3faa9c4d9c44974373dcf5256 is v1.3.0
|
||||
#9e55ace55aad1ada27516038dfbdc66a8a0763db is v1.4.1
|
||||
#83dd62659fc07d5b7fa93b5d1c1879f93509c7db is v1.4.1 latest
|
||||
for onnx_version in "5af210ca8a1c73aa6bae8754c9346ec54d0a756e" "bae6333e149a59a3faa9c4d9c44974373dcf5256" "9e55ace55aad1ada27516038dfbdc66a8a0763db" "83dd62659fc07d5b7fa93b5d1c1879f93509c7db"; do
|
||||
#0e8d2bc5e51455c70ef790b9f65aa632ed9bc8a7 is v1.4.1 latest
|
||||
for onnx_version in "5af210ca8a1c73aa6bae8754c9346ec54d0a756e" "bae6333e149a59a3faa9c4d9c44974373dcf5256" "9e55ace55aad1ada27516038dfbdc66a8a0763db" "0e8d2bc5e51455c70ef790b9f65aa632ed9bc8a7"; do
|
||||
if [ -z ${lastest_onnx_version+x} ]; then
|
||||
echo "first pass";
|
||||
else
|
||||
|
|
|
|||
Loading…
Reference in a new issue