Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 1 | /* |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 2 | * Copyright (c) 2018-2019 ARM Limited. |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 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_WIDTHCONCATENATE_LAYER_FIXTURE |
| 25 | #define ARM_COMPUTE_TEST_WIDTHCONCATENATE_LAYER_FIXTURE |
| 26 | |
| 27 | #include "arm_compute/core/TensorShape.h" |
| 28 | #include "arm_compute/core/Types.h" |
| 29 | #include "arm_compute/core/utils/misc/ShapeCalculator.h" |
| 30 | #include "tests/AssetsLibrary.h" |
| 31 | #include "tests/Globals.h" |
| 32 | #include "tests/IAccessor.h" |
| 33 | #include "tests/framework/Asserts.h" |
| 34 | #include "tests/framework/Fixture.h" |
| 35 | #include "tests/validation/Helpers.h" |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 36 | #include "tests/validation/reference/ConcatenateLayer.h" |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 37 | |
| 38 | #include <random> |
| 39 | |
| 40 | namespace arm_compute |
| 41 | { |
| 42 | namespace test |
| 43 | { |
| 44 | namespace validation |
| 45 | { |
| 46 | template <typename TensorType, typename ITensorType, typename AccessorType, typename FunctionType, typename T> |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 47 | class ConcatenateLayerValidationFixture : public framework::Fixture |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 48 | { |
| 49 | public: |
| 50 | template <typename...> |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 51 | void setup(TensorShape shape, DataType data_type, unsigned int axis) |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 52 | { |
| 53 | // Create input shapes |
| 54 | std::mt19937 gen(library->seed()); |
Michele Di Giorgio | 27400b9 | 2018-11-01 13:44:05 +0000 | [diff] [blame] | 55 | std::uniform_int_distribution<> num_dis(2, 8); |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 56 | std::uniform_int_distribution<> offset_dis(0, 20); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 57 | |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 58 | const int num_tensors = num_dis(gen); |
| 59 | |
| 60 | std::vector<TensorShape> shapes(num_tensors, shape); |
| 61 | |
| 62 | // vector holding the quantization info: |
| 63 | // the last element is the output quantization info |
| 64 | // all other elements are the quantization info for the input tensors |
| 65 | std::vector<QuantizationInfo> qinfo(num_tensors + 1, QuantizationInfo()); |
| 66 | for(auto &qi : qinfo) |
| 67 | { |
| 68 | qi = QuantizationInfo(1.f / 255.f, offset_dis(gen)); |
| 69 | } |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 70 | std::bernoulli_distribution mutate_dis(0.5f); |
| 71 | std::uniform_real_distribution<> change_dis(-0.25f, 0.f); |
| 72 | |
| 73 | // Generate more shapes based on the input |
| 74 | for(auto &s : shapes) |
| 75 | { |
Michalis Spyrou | a9c4472 | 2019-04-05 17:18:36 +0100 | [diff] [blame] | 76 | // Randomly change the dimension |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 77 | if(mutate_dis(gen)) |
| 78 | { |
| 79 | // Decrease the dimension by a small percentage. Don't increase |
| 80 | // as that could make tensor too large. |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 81 | s.set(axis, s[axis] + 2 * static_cast<int>(s[axis] * change_dis(gen))); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 82 | } |
| 83 | } |
| 84 | |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 85 | _target = compute_target(shapes, qinfo, data_type, axis); |
| 86 | _reference = compute_reference(shapes, qinfo, data_type, axis); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 87 | } |
| 88 | |
| 89 | protected: |
| 90 | template <typename U> |
| 91 | void fill(U &&tensor, int i) |
| 92 | { |
| 93 | library->fill_tensor_uniform(tensor, i); |
| 94 | } |
| 95 | |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 96 | TensorType compute_target(const std::vector<TensorShape> &shapes, const std::vector<QuantizationInfo> &qinfo, DataType data_type, unsigned int axis) |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 97 | { |
| 98 | std::vector<TensorType> srcs; |
| 99 | std::vector<ITensorType *> src_ptrs; |
| 100 | |
| 101 | // Create tensors |
| 102 | srcs.reserve(shapes.size()); |
| 103 | |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 104 | for(size_t j = 0; j < shapes.size(); ++j) |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 105 | { |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 106 | srcs.emplace_back(create_tensor<TensorType>(shapes[j], data_type, 1, qinfo[j])); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 107 | src_ptrs.emplace_back(&srcs.back()); |
| 108 | } |
| 109 | |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 110 | const TensorShape dst_shape = misc::shape_calculator::calculate_concatenate_shape(src_ptrs, axis); |
| 111 | TensorType dst = create_tensor<TensorType>(dst_shape, data_type, 1, qinfo[shapes.size()]); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 112 | |
| 113 | // Create and configure function |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 114 | FunctionType concat; |
Georgios Pinitas | 9e4824c | 2019-04-12 13:15:58 +0100 | [diff] [blame] | 115 | concat.configure(src_ptrs, &dst, axis); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 116 | |
| 117 | for(auto &src : srcs) |
| 118 | { |
| 119 | ARM_COMPUTE_EXPECT(src.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 120 | } |
| 121 | |
| 122 | ARM_COMPUTE_EXPECT(dst.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 123 | |
| 124 | // Allocate tensors |
| 125 | for(auto &src : srcs) |
| 126 | { |
| 127 | src.allocator()->allocate(); |
| 128 | ARM_COMPUTE_EXPECT(!src.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 129 | } |
| 130 | |
| 131 | dst.allocator()->allocate(); |
| 132 | ARM_COMPUTE_EXPECT(!dst.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 133 | |
| 134 | // Fill tensors |
| 135 | int i = 0; |
| 136 | for(auto &src : srcs) |
| 137 | { |
| 138 | fill(AccessorType(src), i++); |
| 139 | } |
| 140 | |
| 141 | // Compute function |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 142 | concat.run(); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 143 | |
| 144 | return dst; |
| 145 | } |
| 146 | |
Michalis Spyrou | a9c4472 | 2019-04-05 17:18:36 +0100 | [diff] [blame] | 147 | SimpleTensor<T> compute_reference(std::vector<TensorShape> &shapes, const std::vector<QuantizationInfo> &qinfo, DataType data_type, unsigned int axis) |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 148 | { |
| 149 | std::vector<SimpleTensor<T>> srcs; |
Michalis Spyrou | a9c4472 | 2019-04-05 17:18:36 +0100 | [diff] [blame] | 150 | std::vector<TensorShape *> src_ptrs; |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 151 | |
| 152 | // Create and fill tensors |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 153 | for(size_t j = 0; j < shapes.size(); ++j) |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 154 | { |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 155 | srcs.emplace_back(shapes[j], data_type, 1, qinfo[j]); |
| 156 | fill(srcs.back(), j); |
Michalis Spyrou | a9c4472 | 2019-04-05 17:18:36 +0100 | [diff] [blame] | 157 | src_ptrs.emplace_back(&shapes[j]); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 158 | } |
| 159 | |
Michalis Spyrou | a9c4472 | 2019-04-05 17:18:36 +0100 | [diff] [blame] | 160 | const TensorShape dst_shape = misc::shape_calculator::calculate_concatenate_shape(src_ptrs, axis); |
Pablo Tello | 54e98d9 | 2019-02-05 16:16:19 +0000 | [diff] [blame] | 161 | SimpleTensor<T> dst{ dst_shape, data_type, 1, qinfo[shapes.size()] }; |
Pablo Tello | 3dd5b68 | 2019-03-04 14:14:02 +0000 | [diff] [blame] | 162 | return reference::concatenate_layer<T>(srcs, dst, axis); |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 163 | } |
| 164 | |
| 165 | TensorType _target{}; |
| 166 | SimpleTensor<T> _reference{}; |
Michalis Spyrou | 55b3d12 | 2018-05-09 09:59:23 +0100 | [diff] [blame] | 167 | }; |
| 168 | } // namespace validation |
| 169 | } // namespace test |
| 170 | } // namespace arm_compute |
| 171 | #endif /* ARM_COMPUTE_TEST_WIDTHCONCATENATE_LAYER_FIXTURE */ |