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
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
|
/*
* Copyright (c) 2019 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/core/CL/kernels/CLDepthwiseConvolutionLayerNativeKernel.h"
#include "arm_compute/core/KernelDescriptors.h"
#include "arm_compute/core/Types.h"
#include "arm_compute/core/utils/misc/ShapeCalculator.h"
#include "arm_compute/runtime/CL/CLTensor.h"
#include "arm_compute/runtime/CL/CLTensorAllocator.h"
#include "tests/CL/CLAccessor.h"
#include "tests/CL/Helper.h"
#include "tests/PaddingCalculator.h"
#include "tests/datasets/ShapeDatasets.h"
#include "tests/framework/Asserts.h"
#include "tests/framework/Macros.h"
#include "tests/framework/datasets/Datasets.h"
#include "tests/validation/Validation.h"
#include "tests/validation/fixtures/DepthwiseConvolutionLayerFixture.h"
namespace arm_compute
{
namespace test
{
namespace validation
{
using namespace arm_compute::misc::shape_calculator;
// Create function for CLDepthwiseConvolutionLayerNativeKernel
using CLDepthwiseConvolutionLayerNative = CLSynthetizeFunction<CLDepthwiseConvolutionLayerNativeKernel>;
// Fixture for CLDepthwiseConvolutionLayerNative
template <typename T>
using CLDepthwiseConvolutionLayerNativeFixture = DepthwiseConvolutionLayerNativeConfigurableValidationFixture<CLTensor, CLAccessor, CLDepthwiseConvolutionLayerNative, T>;
namespace
{
// *INDENT-OFF*
// clang-format off
RelativeTolerance<float> rel_tolerance_f32(0.001f);
constexpr float abs_tolerance_f32(0.0001f);
RelativeTolerance<half_float::half> tolerance_f16(half_float::half(0.01));
/** Width values to test - Precommit */
const auto width_values_precommit = framework::dataset::make("width", { 37U } );
/** Width values to test - Nightly */
const auto width_values_nightly = framework::dataset::make("width", { 53U, 47U } );
/** Height values to test - Precommit */
const auto height_values_precommit = framework::dataset::make("height", { 19U } );
/** Height values to test - Nightly */
const auto height_values_nightly = framework::dataset::make("height", { 39U, 43U } );
/** Channel values to test - Precommit */
const auto channel_values_precommit = framework::dataset::make("channels", { 15U });
/** Channel values to test - Nightly */
const auto channel_values_nightly = framework::dataset::make("channels", { 33U, 19U });
/** Batch values to test - Precommit */
const auto batch_values_precommit = framework::dataset::make("batch", { 1U, 2U });
/** Batch values to test - Nightly */
const auto batch_values_nightly = framework::dataset::make("batch", { 1U, 3U });
/** Kernel size values to test - Precommit */
const auto kernel_sz_values_precommit = framework::dataset::make("kernel_size", { Size2D(1U, 1U), Size2D(1U, 3U), Size2D(5U, 5U) });
/** Kernel size values to test - Nightly */
const auto kernel_sz_values_nightly = framework::dataset::make("kernel_size", { Size2D(3U, 5U), Size2D(5U, 1U), Size2D(1U, 7U), Size2D(9U, 7U) });
/** Depth multiplier values to test - All */
const auto depth_multiplier_values = framework::dataset::make("depth_multiplier", {3U});
/** Dilation values to test - All */
const auto dilation_values = framework::dataset::make("dilation", { Size2D(1U, 1U), Size2D(3U, 3U) });
/** Stride values to test - All */
const auto stride_values = framework::dataset::make("stride", { Size2D(1U, 1U), Size2D(3U, 2U) });
/** Padding values to test - All */
const auto padding_valid_values = framework::dataset::make("padding_valid", { true, false });
/** Data type values to test - All */
const auto data_type_values = framework::dataset::make("data_type", { DataType::F32, DataType::F16 });
/** Data layout values to test - All */
const auto data_layout_values = framework::dataset::make("data_layout", { DataLayout::NHWC });
/** N0 values to test - Precommit */
const auto n0_values_precommit = framework::dataset::make("N0", {2, 4});
/** N0 values to test - Nightly */
const auto n0_values_nightly = framework::dataset::make("N0", {3, 8});
/** Activation values to test */
const auto act_values = framework::dataset::make("Activation",
{
ActivationLayerInfo(),
ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, 8.f, 2.f),
});
} // namespace
TEST_SUITE(CL)
TEST_SUITE(DepthwiseConvolutionLayerNative)
TEST_SUITE(Float)
TEST_SUITE(FP32)
FIXTURE_DATA_TEST_CASE(RunSmall, CLDepthwiseConvolutionLayerNativeFixture<float>, framework::DatasetMode::ALL,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_precommit,
height_values_precommit),
channel_values_precommit),
batch_values_precommit),
kernel_sz_values_precommit),
framework::dataset::make("depth_multiplier", 1)),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F32)),
data_layout_values),
act_values),
n0_values_precommit))
{
// Validate output
validate(CLAccessor(_target), _reference, rel_tolerance_f32, 0.f, abs_tolerance_f32);
}
FIXTURE_DATA_TEST_CASE(RunLarge, CLDepthwiseConvolutionLayerNativeFixture<float>, framework::DatasetMode::NIGHTLY,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_nightly,
height_values_nightly),
channel_values_nightly),
batch_values_nightly),
kernel_sz_values_nightly),
framework::dataset::make("depth_multiplier", 1)),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F32)),
data_layout_values),
act_values),
n0_values_nightly))
{
// Validate output
validate(CLAccessor(_target), _reference, rel_tolerance_f32, 0.f, abs_tolerance_f32);
}
TEST_SUITE_END() // FP32
TEST_SUITE(FP16)
FIXTURE_DATA_TEST_CASE(RunSmall, CLDepthwiseConvolutionLayerNativeFixture<half>, framework::DatasetMode::ALL,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_precommit,
height_values_precommit),
channel_values_precommit),
batch_values_precommit),
kernel_sz_values_precommit),
framework::dataset::make("depth_multiplier", 1)),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F16)),
data_layout_values),
act_values),
n0_values_precommit))
{
// Validate output
validate(CLAccessor(_target), _reference, tolerance_f16);
}
FIXTURE_DATA_TEST_CASE(RunLarge, CLDepthwiseConvolutionLayerNativeFixture<half>, framework::DatasetMode::NIGHTLY,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_nightly,
height_values_nightly),
channel_values_nightly),
batch_values_nightly),
kernel_sz_values_nightly),
framework::dataset::make("depth_multiplier", 1)),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F16)),
data_layout_values),
act_values),
n0_values_nightly))
{
// Validate output
validate(CLAccessor(_target), _reference, tolerance_f16);
}
TEST_SUITE_END() // FP16
TEST_SUITE_END() // Float
TEST_SUITE(DepthMultiplier)
TEST_SUITE(Float)
TEST_SUITE(FP32)
FIXTURE_DATA_TEST_CASE(RunSmall, CLDepthwiseConvolutionLayerNativeFixture<float>, framework::DatasetMode::ALL,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_precommit,
height_values_precommit),
channel_values_precommit),
batch_values_precommit),
kernel_sz_values_precommit),
depth_multiplier_values),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F32)),
data_layout_values),
act_values),
framework::dataset::make("N0", 1)))
{
// Validate output
validate(CLAccessor(_target), _reference, rel_tolerance_f32, 0.f, abs_tolerance_f32);
}
FIXTURE_DATA_TEST_CASE(RunLarge, CLDepthwiseConvolutionLayerNativeFixture<float>, framework::DatasetMode::NIGHTLY,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_nightly,
height_values_nightly),
channel_values_nightly),
batch_values_nightly),
kernel_sz_values_nightly),
depth_multiplier_values),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F32)),
data_layout_values),
act_values),
framework::dataset::make("N0", 1)))
{
// Validate output
validate(CLAccessor(_target), _reference, rel_tolerance_f32, 0.f, abs_tolerance_f32);
}
TEST_SUITE_END() // FP32
TEST_SUITE(FP16)
FIXTURE_DATA_TEST_CASE(RunSmall, CLDepthwiseConvolutionLayerNativeFixture<half>, framework::DatasetMode::ALL,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_precommit,
height_values_precommit),
channel_values_precommit),
batch_values_precommit),
kernel_sz_values_precommit),
depth_multiplier_values),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F16)),
data_layout_values),
act_values),
framework::dataset::make("N0", 1)))
{
// Validate output
validate(CLAccessor(_target), _reference, tolerance_f16);
}
FIXTURE_DATA_TEST_CASE(RunLarge, CLDepthwiseConvolutionLayerNativeFixture<half>, framework::DatasetMode::NIGHTLY,
combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(combine(
width_values_nightly,
height_values_nightly),
channel_values_nightly),
batch_values_nightly),
kernel_sz_values_nightly),
depth_multiplier_values),
dilation_values),
stride_values),
padding_valid_values),
framework::dataset::make("DataType", DataType::F16)),
data_layout_values),
act_values),
framework::dataset::make("N0", 1)))
{
// Validate output
validate(CLAccessor(_target), _reference, tolerance_f16);
}
TEST_SUITE_END() // FP16
TEST_SUITE_END() // Float
TEST_SUITE_END() // DepthMultiplier
TEST_SUITE_END() // DepthwiseConvolutionLayerNative
TEST_SUITE_END() // CL
} // namespace validation
} // namespace test
} // namespace arm_compute
|
