blob: e0921ed7ce7e4becc2282f26e3717f35a9a18169 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
|
//
// Copyright © 2017 Arm Ltd. All rights reserved.
// SPDX-License-Identifier: MIT
//
#pragma once
#include <armnn/Tensor.hpp>
#include <armnn/Descriptors.hpp>
#include <armnn/backends/Workload.hpp>
#include <arm_compute/runtime/CL/functions/CLConvolutionLayer.h>
#include <arm_compute/runtime/MemoryManagerOnDemand.h>
#include <memory>
namespace armnn
{
arm_compute::Status ClConvolution2dWorkloadValidate(const TensorInfo& input,
const TensorInfo& output,
const Convolution2dDescriptor& descriptor,
const TensorInfo& weights,
const Optional<TensorInfo>& biases,
bool isFastMathEnabled = false,
const ActivationDescriptor* activationDescriptor = nullptr);
class ClConvolution2dWorkload : public BaseWorkload<Convolution2dQueueDescriptor>
{
public:
ClConvolution2dWorkload(const Convolution2dQueueDescriptor& descriptor,
const WorkloadInfo& info,
std::shared_ptr<arm_compute::MemoryManagerOnDemand>& memoryManager,
const arm_compute::CLCompileContext& clCompileContext,
const bool isFastMathEnabled = false);
void Execute() const override;
arm_compute::ConvolutionMethod GetConvolutionMethod() const;
private:
mutable arm_compute::CLConvolutionLayer m_ConvolutionLayer;
std::unique_ptr<arm_compute::CLTensor> m_KernelTensor;
std::unique_ptr<arm_compute::CLTensor> m_BiasTensor;
arm_compute::ConvolutionMethod m_ConvolutionMethod;
void FreeUnusedTensors();
};
} //namespace armnn
|
