• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 /*
2  * Copyright (c) 2017-2019 Arm Limited.
3  *
4  * SPDX-License-Identifier: MIT
5  *
6  * Permission is hereby granted, free of charge, to any person obtaining a copy
7  * of this software and associated documentation files (the "Software"), to
8  * deal in the Software without restriction, including without limitation the
9  * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
10  * sell copies of the Software, and to permit persons to whom the Software is
11  * furnished to do so, subject to the following conditions:
12  *
13  * The above copyright notice and this permission notice shall be included in all
14  * copies or substantial portions of the Software.
15  *
16  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
17  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
18  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
19  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
20  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
21  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
22  * SOFTWARE.
23  */
24 #ifndef ARM_COMPUTE_TEST_NORMALIZE_PLANAR_YUV_LAYER_FIXTURE
25 #define ARM_COMPUTE_TEST_NORMALIZE_PLANAR_YUV_LAYER_FIXTURE
26 
27 #include "arm_compute/core/TensorShape.h"
28 #include "arm_compute/core/Types.h"
29 #include "tests/AssetsLibrary.h"
30 #include "tests/Globals.h"
31 #include "tests/IAccessor.h"
32 #include "tests/framework/Asserts.h"
33 #include "tests/framework/Fixture.h"
34 #include "tests/validation/Helpers.h"
35 #include "tests/validation/reference/NormalizePlanarYUVLayer.h"
36 
37 namespace arm_compute
38 {
39 namespace test
40 {
41 namespace validation
42 {
43 template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
44 class NormalizePlanarYUVLayerValidationGenericFixture : public framework::Fixture
45 {
46 public:
47     template <typename...>
setup(TensorShape shape0,TensorShape shape1,DataType dt,DataLayout data_layout,QuantizationInfo quantization_info)48     void setup(TensorShape shape0, TensorShape shape1, DataType dt, DataLayout data_layout, QuantizationInfo quantization_info)
49     {
50         _data_type = dt;
51         _target    = compute_target(shape0, shape1, dt, data_layout, quantization_info);
52         _reference = compute_reference(shape0, shape1, dt, quantization_info);
53     }
54 
55 protected:
56     template <typename U>
fill(U && src_tensor,U && mean_tensor,U && std_tensor)57     void fill(U &&src_tensor, U &&mean_tensor, U &&std_tensor)
58     {
59         if(is_data_type_float(_data_type))
60         {
61             const float                      min_bound = -1.f;
62             const float                      max_bound = 1.f;
63             std::uniform_real_distribution<> distribution(min_bound, max_bound);
64             std::uniform_real_distribution<> distribution_std(0.1, max_bound);
65             library->fill(src_tensor, distribution, 0);
66             library->fill(mean_tensor, distribution, 1);
67             library->fill(std_tensor, distribution_std, 2);
68         }
69         else if(is_data_type_quantized_asymmetric(_data_type))
70         {
71             const QuantizationInfo quant_info = src_tensor.quantization_info();
72             std::pair<int, int> bounds = get_quantized_bounds(quant_info, -1.f, 1.0f);
73             std::uniform_int_distribution<> distribution(bounds.first, bounds.second);
74             std::uniform_int_distribution<> distribution_std(quantize_qasymm8(0.1f, quant_info.uniform()), bounds.second);
75             library->fill(src_tensor, distribution, 0);
76             library->fill(mean_tensor, distribution, 1);
77             library->fill(std_tensor, distribution_std, 2);
78         }
79     }
80 
compute_target(TensorShape shape0,const TensorShape & shape1,DataType dt,DataLayout data_layout,QuantizationInfo quantization_info)81     TensorType compute_target(TensorShape shape0, const TensorShape &shape1, DataType dt, DataLayout data_layout, QuantizationInfo quantization_info)
82     {
83         if(data_layout == DataLayout::NHWC)
84         {
85             permute(shape0, PermutationVector(2U, 0U, 1U));
86         }
87 
88         // Create tensors
89         TensorType src  = create_tensor<TensorType>(shape0, dt, 1, quantization_info, data_layout);
90         TensorType mean = create_tensor<TensorType>(shape1, dt, 1, quantization_info);
91         TensorType std  = create_tensor<TensorType>(shape1, dt, 1, quantization_info);
92         TensorType dst;
93 
94         // Create and configure function
95         FunctionType norm;
96         norm.configure(&src, &dst, &mean, &std);
97 
98         ARM_COMPUTE_EXPECT(src.info()->is_resizable(), framework::LogLevel::ERRORS);
99         ARM_COMPUTE_EXPECT(dst.info()->is_resizable(), framework::LogLevel::ERRORS);
100         ARM_COMPUTE_EXPECT(mean.info()->is_resizable(), framework::LogLevel::ERRORS);
101         ARM_COMPUTE_EXPECT(std.info()->is_resizable(), framework::LogLevel::ERRORS);
102 
103         // Allocate tensors
104         src.allocator()->allocate();
105         dst.allocator()->allocate();
106         mean.allocator()->allocate();
107         std.allocator()->allocate();
108 
109         ARM_COMPUTE_EXPECT(!src.info()->is_resizable(), framework::LogLevel::ERRORS);
110         ARM_COMPUTE_EXPECT(!dst.info()->is_resizable(), framework::LogLevel::ERRORS);
111         ARM_COMPUTE_EXPECT(!mean.info()->is_resizable(), framework::LogLevel::ERRORS);
112         ARM_COMPUTE_EXPECT(!std.info()->is_resizable(), framework::LogLevel::ERRORS);
113 
114         // Fill tensors
115         fill(AccessorType(src), AccessorType(mean), AccessorType(std));
116 
117         // Compute function
118         norm.run();
119 
120         return dst;
121     }
122 
compute_reference(const TensorShape & shape0,const TensorShape & shape1,DataType dt,QuantizationInfo quantization_info)123     SimpleTensor<T> compute_reference(const TensorShape &shape0, const TensorShape &shape1, DataType dt, QuantizationInfo quantization_info)
124     {
125         // Create reference
126         SimpleTensor<T> ref_src{ shape0, dt, 1, quantization_info };
127         SimpleTensor<T> ref_mean{ shape1, dt, 1, quantization_info };
128         SimpleTensor<T> ref_std{ shape1, dt, 1, quantization_info };
129 
130         // Fill reference
131         fill(ref_src, ref_mean, ref_std);
132 
133         return reference::normalize_planar_yuv_layer(ref_src, ref_mean, ref_std);
134     }
135 
136     TensorType      _target{};
137     SimpleTensor<T> _reference{};
138     DataType        _data_type{};
139 };
140 
141 template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
142 class NormalizePlanarYUVLayerValidationFixture : public NormalizePlanarYUVLayerValidationGenericFixture<TensorType, AccessorType, FunctionType, T>
143 {
144 public:
145     template <typename...>
setup(TensorShape shape0,TensorShape shape1,DataType dt,DataLayout data_layout)146     void setup(TensorShape shape0, TensorShape shape1, DataType dt, DataLayout data_layout)
147     {
148         NormalizePlanarYUVLayerValidationGenericFixture<TensorType, AccessorType, FunctionType, T>::setup(shape0, shape1, dt, data_layout, QuantizationInfo());
149     }
150 };
151 
152 template <typename TensorType, typename AccessorType, typename FunctionType, typename T>
153 class NormalizePlanarYUVLayerValidationQuantizedFixture : public NormalizePlanarYUVLayerValidationGenericFixture<TensorType, AccessorType, FunctionType, T>
154 {
155 public:
156     template <typename...>
setup(TensorShape shape0,TensorShape shape1,DataType dt,DataLayout data_layout,QuantizationInfo quantization_info)157     void setup(TensorShape shape0, TensorShape shape1, DataType dt, DataLayout data_layout, QuantizationInfo quantization_info)
158     {
159         NormalizePlanarYUVLayerValidationGenericFixture<TensorType, AccessorType, FunctionType, T>::setup(shape0, shape1, dt, data_layout, quantization_info);
160     }
161 };
162 } // namespace validation
163 } // namespace test
164 } // namespace arm_compute
165 #endif /* ARM_COMPUTE_TEST_NORMALIZE_PLANAR_YUV_LAYER_FIXTURE */
166