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
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
|
/*
* Copyright (c) 2017-2018 ARM Limited.
*
* SPDX-License-Identifier: MIT
*
* 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 "arm_compute/runtime/NEON/functions/NEConvolutionLayer.h"
#include "arm_compute/core/PixelValue.h"
#include "arm_compute/core/Utils.h"
#include "arm_compute/core/Validate.h"
#include "support/ToolchainSupport.h"
#include <cmath>
#include <tuple>
#include <utility>
namespace arm_compute
{
NEConvolutionLayer::NEConvolutionLayer(std::shared_ptr<IMemoryManager> memory_manager) //NOLINT
: _memory_manager(std::move(memory_manager)),
_function()
{
}
void NEConvolutionLayer::configure(ITensor *input, const ITensor *weights, const ITensor *biases, ITensor *output, const PadStrideInfo &conv_info, const WeightsInfo &weights_info,
const Size2D &dilation, const ActivationLayerInfo &act_info, bool enable_fast_math)
{
// Perform validate step
ARM_COMPUTE_ERROR_ON_NULLPTR(input, weights, output);
ARM_COMPUTE_ERROR_THROW_ON(NEConvolutionLayer::validate(input->info(), weights->info(), ((biases != nullptr) ? biases->info() : nullptr), output->info(), conv_info, weights_info, dilation, act_info,
enable_fast_math));
switch(NEConvolutionLayer::get_convolution_method(input->info(), weights->info(), output->info(), conv_info, weights_info, dilation, act_info))
{
case ConvolutionMethod::WINOGRAD:
{
auto f = arm_compute::support::cpp14::make_unique<NEWinogradConvolutionLayer>(_memory_manager);
f->configure(input, weights, biases, output, conv_info, act_info, enable_fast_math);
_function = std::move(f);
break;
}
case ConvolutionMethod::GEMM:
{
auto f = arm_compute::support::cpp14::make_unique<NEGEMMConvolutionLayer>(_memory_manager);
f->configure(input, weights, biases, output, conv_info, weights_info, dilation, act_info);
_function = std::move(f);
break;
}
case ConvolutionMethod::DIRECT:
{
auto f = arm_compute::support::cpp14::make_unique<NEDirectConvolutionLayer>(_memory_manager);
f->configure(input, weights, biases, output, conv_info, act_info);
_function = std::move(f);
break;
}
default:
ARM_COMPUTE_ERROR("Not supported.");
break;
}
}
Status NEConvolutionLayer::validate(const ITensorInfo *input, const ITensorInfo *weights, const ITensorInfo *biases, const ITensorInfo *output, const PadStrideInfo &conv_info,
const WeightsInfo &weights_info, const Size2D &dilation, const ActivationLayerInfo &act_info, bool enable_fast_math)
{
switch(NEConvolutionLayer::get_convolution_method(input, weights, output, conv_info, weights_info, dilation, act_info))
{
case ConvolutionMethod::WINOGRAD:
//Validate Winograd
NEWinogradConvolutionLayer::validate(input, weights, biases, output, conv_info, act_info, enable_fast_math);
break;
case ConvolutionMethod::GEMM:
//Validate Gemm-based Convolution
NEGEMMConvolutionLayer::validate(input, weights, biases, output, conv_info, weights_info, dilation, act_info);
break;
case ConvolutionMethod::DIRECT:
//Validate Gemm-based Convolution
NEDirectConvolutionLayer::validate(input, weights, biases, output, conv_info, act_info);
default:
ARM_COMPUTE_ERROR("Not supported.");
break;
}
return Status{};
}
ConvolutionMethod NEConvolutionLayer::get_convolution_method(const ITensorInfo *input, const ITensorInfo *weights,
const ITensorInfo *output, const PadStrideInfo &conv_info,
const WeightsInfo &weights_info, const Size2D &dilation, const ActivationLayerInfo &act_info, bool enable_fast_math)
{
ARM_COMPUTE_ERROR_ON_NULLPTR(input, output, weights);
ARM_COMPUTE_UNUSED(weights_info);
if(dilation != Size2D(1U, 1U) || Scheduler::get().cpu_info().get_cpu_model() == CPUModel::A53
|| input->dimension(get_data_layout_dimension_index(input->data_layout(), DataLayoutDimension::CHANNEL)) <= 16)
{
return ConvolutionMethod::GEMM;
}
return bool(NEWinogradConvolutionLayer::validate(input, weights, nullptr, output, conv_info, act_info, enable_fast_math)) ? ConvolutionMethod::WINOGRAD : ConvolutionMethod::GEMM;
}
void NEConvolutionLayer::run()
{
_function->run();
}
} // namespace arm_compute
|
