Added topk & unit test (#384)

Signed-off-by: Chen Xin <jack.chen@verisilicon.com>

Co-authored-by: Chen Xin <jack.chen@verisilicon.com>
This commit is contained in:
chxin66 2022-05-05 17:06:39 +08:00 committed by GitHub
parent c09cdf79ad
commit 7a8ae32f73
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 210 additions and 0 deletions

55
include/tim/vx/ops/topk.h Normal file
View File

@ -0,0 +1,55 @@
/****************************************************************************
*
* Copyright (c) 2022 Vivante Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS IN THE SOFTWARE.
*
*****************************************************************************/
#ifndef TIM_VX_OPS_TOPK_H_
#define TIM_VX_OPS_TOPK_H_
#include <array>
#include "tim/vx/direct_map_op.h"
#include "tim/vx/types.h"
namespace tim {
namespace vx {
namespace ops {
/**
* ## Topk
*
* Finds values and indices of the k largest entries for the last dimension.
*
* - k : Number of top elements to look for along the last dimension.
*/
class Topk : public DirectMapOp {
public:
Topk(Graph* graph, uint32_t k);
std::shared_ptr<Operation> Clone(std::shared_ptr<Graph>& graph) const override;
};
} // namespace ops
} // namespace vx
} // namespace tim
#endif /* TIM_VX_OPS_TOPK_H_ */

45
src/tim/vx/ops/topk.cc Normal file
View File

@ -0,0 +1,45 @@
/****************************************************************************
*
* Copyright (c) 2022 Vivante Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS IN THE SOFTWARE.
*
*****************************************************************************/
#include "tim/vx/ops/topk.h"
#include "direct_map_op_impl.h"
#include "type_utils.h"
#include "vsi_nn_pub.h"
namespace tim {
namespace vx {
namespace ops {
Topk::Topk(Graph* graph, uint32_t k)
: DirectMapOp(graph, VSI_NN_OP_TOPK) {
this->impl()->node()->nn_param.topk.k = k;
}
std::shared_ptr<Operation> Topk::Clone(std::shared_ptr<Graph>& graph) const {
return graph->CreateOperation<Topk>(this->impl()->node()->nn_param.topk.k);
}
} // namespace ops
} // namespace vx
} // namespace tim

110
src/tim/vx/ops/topk_test.cc Normal file
View File

@ -0,0 +1,110 @@
/****************************************************************************
*
* Copyright (c) 2022 Vivante Corporation
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, sublicense,
* and/or sell copies of the Software, and to permit persons to whom the
* Software is furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
* DEALINGS IN THE SOFTWARE.
*
*****************************************************************************/
#include "tim/vx/context.h"
#include "tim/vx/graph.h"
#include "tim/vx/ops/topk.h"
#include "gtest/gtest.h"
TEST(Topk, shape_3_4_k_2) {
auto ctx = tim::vx::Context::Create();
auto graph = ctx->CreateGraph();
tim::vx::ShapeType input_shape({4, 3});
tim::vx::ShapeType values_shape({6});
tim::vx::TensorSpec input_spec(tim::vx::DataType::FLOAT32, input_shape,
tim::vx::TensorAttribute::INPUT);
tim::vx::TensorSpec values_spec(tim::vx::DataType::FLOAT32, values_shape,
tim::vx::TensorAttribute::OUTPUT);
tim::vx::TensorSpec indices_spec(tim::vx::DataType::INT32, values_shape,
tim::vx::TensorAttribute::OUTPUT);
auto input_tensor = graph->CreateTensor(input_spec);
auto value_tensor = graph->CreateTensor(values_spec);
auto indices_tensor = graph->CreateTensor(indices_spec);
std::vector<float> in_data = {
1, 2, 3, 4, 1, 2, 3, 4, 1, 2, 3, 4,
};
std::vector<float> value_golden = {
4, 3, 4, 3, 4, 3,
};
std::vector<int32_t> indices_golden = {
3, 2, 3, 2, 3, 2,
};
EXPECT_TRUE(input_tensor->CopyDataToTensor(in_data.data(),
in_data.size() * sizeof(float)));
auto op = graph->CreateOperation<tim::vx::ops::Topk>(2);
(*op).BindInputs({input_tensor}).BindOutputs({value_tensor, indices_tensor});
EXPECT_TRUE(graph->Compile());
EXPECT_TRUE(graph->Run());
std::vector<float> value(value_golden.size());
std::vector<int32_t> indices(6);
EXPECT_TRUE(value_tensor->CopyDataFromTensor(value.data()));
EXPECT_TRUE(indices_tensor->CopyDataFromTensor(indices.data()));
EXPECT_EQ(value_golden, value);
EXPECT_EQ(indices_golden, indices);
}
TEST(Topk, shape_3_2_2_k_1) {
auto ctx = tim::vx::Context::Create();
auto graph = ctx->CreateGraph();
tim::vx::ShapeType input_shape({2, 2, 3});
tim::vx::ShapeType values_shape({6});
tim::vx::TensorSpec input_spec(tim::vx::DataType::FLOAT32, input_shape,
tim::vx::TensorAttribute::INPUT);
tim::vx::TensorSpec values_spec(tim::vx::DataType::FLOAT32, values_shape,
tim::vx::TensorAttribute::OUTPUT);
tim::vx::TensorSpec indices_spec(tim::vx::DataType::INT32, values_shape,
tim::vx::TensorAttribute::OUTPUT);
auto input_tensor = graph->CreateTensor(input_spec);
auto value_tensor = graph->CreateTensor(values_spec);
auto indices_tensor = graph->CreateTensor(indices_spec);
std::vector<float> in_data = {
1, 2, 3, 4, 1, 2, 3, 4, 1, 2, 3, 4,
};
std::vector<float> value_golden = {
2, 4, 2, 4, 2, 4,
};
std::vector<int32_t> indices_golden = {
1, 1, 1, 1, 1, 1,
};
EXPECT_TRUE(input_tensor->CopyDataToTensor(in_data.data(),
in_data.size() * sizeof(float)));
auto op = graph->CreateOperation<tim::vx::ops::Topk>(1);
(*op).BindInputs({input_tensor}).BindOutputs({value_tensor, indices_tensor});
EXPECT_TRUE(graph->Compile());
EXPECT_TRUE(graph->Run());
std::vector<float> value(value_golden.size());
std::vector<int32_t> indices(6);
EXPECT_TRUE(value_tensor->CopyDataFromTensor(value.data()));
EXPECT_TRUE(indices_tensor->CopyDataFromTensor(indices.data()));
EXPECT_EQ(value_golden, value);
EXPECT_EQ(indices_golden, indices);
}