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
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
|
/*
* Copyright (c) 2017-2021, 2023-2024 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.
*/
#ifndef ACL_ARM_COMPUTE_RUNTIME_CL_FUNCTIONS_CLDEPTHWISECONVOLUTIONLAYER_H
#define ACL_ARM_COMPUTE_RUNTIME_CL_FUNCTIONS_CLDEPTHWISECONVOLUTIONLAYER_H
#include "arm_compute/core/Types.h"
#include "arm_compute/function_info/ActivationLayerInfo.h"
#include "arm_compute/runtime/CL/CLTensor.h"
#include "arm_compute/runtime/CL/functions/CLPermute.h"
#include "arm_compute/runtime/IFunction.h"
#include "arm_compute/runtime/MemoryGroup.h"
namespace arm_compute
{
class CLCompileContext;
class CLDepthwiseConvolutionLayerNativeKernel;
class ICLTensor;
/** Function to execute a depthwise convolution
*
* -# CLDepthwiseConvolutionLayerNativeKernel
* -# @ref CLPermute (if the data layout is NCHW)
*
*/
class CLDepthwiseConvolutionLayer : public IFunction
{
public:
/** Default constructor */
CLDepthwiseConvolutionLayer(std::shared_ptr<IMemoryManager> memory_manager = nullptr);
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLDepthwiseConvolutionLayer(const CLDepthwiseConvolutionLayer &) = delete;
/** Default move constructor */
CLDepthwiseConvolutionLayer(CLDepthwiseConvolutionLayer &&) = default;
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLDepthwiseConvolutionLayer &operator=(const CLDepthwiseConvolutionLayer &) = delete;
/** Default move assignment operator */
CLDepthwiseConvolutionLayer &operator=(CLDepthwiseConvolutionLayer &&) = default;
/** Default destructor */
~CLDepthwiseConvolutionLayer();
/** Initialize the function's source, destination, weights and convolution information.
*
* Valid data layouts:
* - NHWC
* - NCHW
*
* Valid data type configurations:
* |src0 |src1 |src2 |dst |
* |:--------------|:------------------|:------|:--------------|
* |F16 |F16 |F16 |F16 |
* |F32 |F32 |F32 |F32 |
* |QASYMM8 |QASYMM8 |S32 |QASYMM8 |
* |QASYMM8 |QSYMM8_PER_CHANNEL |S32 |QASYMM8 |
* |QASYMM8_SIGNED |QASYMM8_SIGNED |S32 |QASYMM8_SIGNED |
* |QASYMM8_SIGNED |QSYMM8_PER_CHANNEL |S32 |QASYMM8_SIGNED |
*
* @param[in] compile_context The compile context to be used.
* @param[in, out] input Source tensor. Data type supported: QASYMM8/QASYMM8_SIGNED/FP16/FP32. Data layout supported: NHWC, NCHW
* @param[in] weights Weights tensor. These are 3D tensors with shape [kernel_x, kernel_y, IFM].
* Data type supported: Same as @p input or QASYMM8/QASYMM8_SIGNED/QSYMM8_PER_CHANNEL when @p input is QASYMM8.
* @param[in] biases Biases tensor. A 1D tensor with shape [IFM]. Must be nullptr if not needed.
* Data type supported: Same as @p input, S32 when input is QASYMM8/QASYMM8_SIGNED.
* @param[out] output Destination tensor. Pass in nullptr or @p input for in-place operation. Data type supported: same as @p input.
* @param[in] conv_info Padding and stride information to use for the convolution.
* @param[in] depth_multiplier (Optional) Multiplier to apply to the input's depth in order to retrieve the output's depth. Defaults to 1.
* @param[in] act_info (Optional) Activation layer information in case of a fused activation.
* @param[in] dilation (Optional) Dilation, in elements, across x and y. Defaults to (1, 1).
*
* @note: For in-place support, please check CLDepthwiseConvolutionLayerNativeKernel
*/
void configure(const CLCompileContext &compile_context,
ICLTensor *input,
const ICLTensor *weights,
const ICLTensor *biases,
ICLTensor *output,
const PadStrideInfo &conv_info,
unsigned int depth_multiplier = 1,
ActivationLayerInfo act_info = ActivationLayerInfo(),
const Size2D &dilation = Size2D(1U, 1U));
/** Initialize the function's source, destination, weights and convolution information.
*
* Similar to @ref CLDepthwiseConvolutionLayer::configure()
*/
void configure(ICLTensor *input,
const ICLTensor *weights,
const ICLTensor *biases,
ICLTensor *output,
const PadStrideInfo &conv_info,
unsigned int depth_multiplier = 1,
ActivationLayerInfo act_info = ActivationLayerInfo(),
const Size2D &dilation = Size2D(1U, 1U));
/** Static function to check if given info will lead to a valid configuration of @ref CLDepthwiseConvolutionLayer
*
* Similar to @ref CLDepthwiseConvolutionLayer::configure()
*
* @return a status
*/
static Status validate(const ITensorInfo *input,
const ITensorInfo *weights,
const ITensorInfo *biases,
const ITensorInfo *output,
const PadStrideInfo &conv_info,
unsigned int depth_multiplier = 1,
ActivationLayerInfo act_info = ActivationLayerInfo(),
const Size2D &dilation = Size2D(1U, 1U));
// Inherited methods overriden:
void run() override;
void prepare() override;
void set_memory_group(std::shared_ptr<IMemoryManager> memory_manager)
{
_memory_group = MemoryGroup(std::move(memory_manager));
};
private:
MemoryGroup _memory_group;
std::unique_ptr<CLDepthwiseConvolutionLayerNativeKernel> _dwc_native_kernel;
CLPermute _permute_input_to_nhwc;
CLPermute _permute_weights_to_nhwc;
CLPermute _permute_output_to_nchw;
CLTensor _permuted_input;
CLTensor _permuted_weights;
CLTensor _permuted_output;
CLTensor _output_multipliers;
CLTensor _output_shifts;
const ITensor *_original_weights;
const ITensor *_input;
const ITensor *_output;
bool _needs_permute;
bool _is_prepared;
bool _is_quantized;
};
} // namespace arm_compute
#endif // ACL_ARM_COMPUTE_RUNTIME_CL_FUNCTIONS_CLDEPTHWISECONVOLUTIONLAYER_H
|
