Michalis Spyrou | bcedf51 | 2018-03-22 14:55:08 +0000 | [diff] [blame] | 1 | /* |
Michele Di Giorgio | d9eaf61 | 2020-07-08 11:12:57 +0100 | [diff] [blame] | 2 | * Copyright (c) 2018 Arm Limited. |
Michalis Spyrou | bcedf51 | 2018-03-22 14:55:08 +0000 | [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_LSTM_LAYER_DATASET |
| 25 | #define ARM_COMPUTE_TEST_LSTM_LAYER_DATASET |
| 26 | |
| 27 | #include "utils/TypePrinter.h" |
| 28 | |
| 29 | #include "arm_compute/core/TensorShape.h" |
| 30 | #include "arm_compute/core/Types.h" |
| 31 | |
| 32 | namespace arm_compute |
| 33 | { |
| 34 | namespace test |
| 35 | { |
| 36 | namespace datasets |
| 37 | { |
| 38 | class LSTMLayerDataset |
| 39 | { |
| 40 | public: |
| 41 | using type = std::tuple<TensorShape, TensorShape, TensorShape, TensorShape, TensorShape, TensorShape, TensorShape, ActivationLayerInfo, float, float>; |
| 42 | |
| 43 | struct iterator |
| 44 | { |
| 45 | iterator(std::vector<TensorShape>::const_iterator src_it, |
| 46 | std::vector<TensorShape>::const_iterator input_weights_it, |
| 47 | std::vector<TensorShape>::const_iterator recurrent_weights_it, |
| 48 | std::vector<TensorShape>::const_iterator cells_bias_it, |
| 49 | std::vector<TensorShape>::const_iterator output_cell_it, |
| 50 | std::vector<TensorShape>::const_iterator dst_it, |
| 51 | std::vector<TensorShape>::const_iterator scratch_it, |
| 52 | std::vector<ActivationLayerInfo>::const_iterator infos_it, |
| 53 | std::vector<float>::const_iterator cell_threshold_it, |
| 54 | std::vector<float>::const_iterator projection_threshold_it) |
| 55 | : _src_it{ std::move(src_it) }, |
| 56 | _input_weights_it{ std::move(input_weights_it) }, |
| 57 | _recurrent_weights_it{ std::move(recurrent_weights_it) }, |
| 58 | _cells_bias_it{ std::move(cells_bias_it) }, |
| 59 | _output_cell_it{ std::move(output_cell_it) }, |
| 60 | _dst_it{ std::move(dst_it) }, |
| 61 | _scratch_it{ std::move(scratch_it) }, |
| 62 | _infos_it{ std::move(infos_it) }, |
| 63 | _cell_threshold_it{ std::move(cell_threshold_it) }, |
| 64 | _projection_threshold_it{ std::move(projection_threshold_it) } |
| 65 | { |
| 66 | } |
| 67 | |
| 68 | std::string description() const |
| 69 | { |
| 70 | std::stringstream description; |
| 71 | description << "In=" << *_src_it << ":"; |
| 72 | description << "InputWeights=" << *_input_weights_it << ":"; |
| 73 | description << "RecurrentWeights=" << *_recurrent_weights_it << ":"; |
| 74 | description << "Biases=" << *_cells_bias_it << ":"; |
| 75 | description << "Scratch=" << *_scratch_it << ":"; |
| 76 | description << "Out=" << *_dst_it; |
| 77 | return description.str(); |
| 78 | } |
| 79 | |
| 80 | LSTMLayerDataset::type operator*() const |
| 81 | { |
| 82 | return std::make_tuple(*_src_it, *_input_weights_it, *_recurrent_weights_it, *_cells_bias_it, *_output_cell_it, *_dst_it, *_scratch_it, *_infos_it, *_cell_threshold_it, *_projection_threshold_it); |
| 83 | } |
| 84 | |
| 85 | iterator &operator++() |
| 86 | { |
| 87 | ++_src_it; |
| 88 | ++_input_weights_it; |
| 89 | ++_recurrent_weights_it; |
| 90 | ++_cells_bias_it; |
| 91 | ++_output_cell_it; |
| 92 | ++_dst_it; |
| 93 | ++_scratch_it; |
| 94 | ++_infos_it; |
| 95 | ++_cell_threshold_it; |
| 96 | ++_projection_threshold_it; |
| 97 | |
| 98 | return *this; |
| 99 | } |
| 100 | |
| 101 | private: |
| 102 | std::vector<TensorShape>::const_iterator _src_it; |
| 103 | std::vector<TensorShape>::const_iterator _input_weights_it; |
| 104 | std::vector<TensorShape>::const_iterator _recurrent_weights_it; |
| 105 | std::vector<TensorShape>::const_iterator _cells_bias_it; |
| 106 | std::vector<TensorShape>::const_iterator _output_cell_it; |
| 107 | std::vector<TensorShape>::const_iterator _dst_it; |
| 108 | std::vector<TensorShape>::const_iterator _scratch_it; |
| 109 | std::vector<ActivationLayerInfo>::const_iterator _infos_it; |
| 110 | std::vector<float>::const_iterator _cell_threshold_it; |
| 111 | std::vector<float>::const_iterator _projection_threshold_it; |
| 112 | }; |
| 113 | |
| 114 | iterator begin() const |
| 115 | { |
| 116 | return iterator(_src_shapes.begin(), _input_weights_shapes.begin(), _recurrent_weights_shapes.begin(), _cell_bias_shapes.begin(), _output_cell_shapes.begin(), _dst_shapes.begin(), |
| 117 | _scratch_shapes.begin(), _infos.begin(), _cell_threshold.begin(), _projection_threshold.begin()); |
| 118 | } |
| 119 | |
| 120 | int size() const |
| 121 | { |
| 122 | return std::min(_src_shapes.size(), std::min(_input_weights_shapes.size(), std::min(_recurrent_weights_shapes.size(), std::min(_cell_bias_shapes.size(), std::min(_output_cell_shapes.size(), |
| 123 | std::min(_dst_shapes.size(), std::min(_scratch_shapes.size(), std::min(_cell_threshold.size(), std::min(_projection_threshold.size(), _infos.size()))))))))); |
| 124 | } |
| 125 | |
| 126 | void add_config(TensorShape src, TensorShape input_weights, TensorShape recurrent_weights, TensorShape cell_bias_weights, TensorShape output_cell_state, TensorShape dst, TensorShape scratch, |
| 127 | ActivationLayerInfo info, float cell_threshold, float projection_threshold) |
| 128 | { |
| 129 | _src_shapes.emplace_back(std::move(src)); |
| 130 | _input_weights_shapes.emplace_back(std::move(input_weights)); |
| 131 | _recurrent_weights_shapes.emplace_back(std::move(recurrent_weights)); |
| 132 | _cell_bias_shapes.emplace_back(std::move(cell_bias_weights)); |
| 133 | _output_cell_shapes.emplace_back(std::move(output_cell_state)); |
| 134 | _dst_shapes.emplace_back(std::move(dst)); |
| 135 | _scratch_shapes.emplace_back(std::move(scratch)); |
| 136 | _infos.emplace_back(std::move(info)); |
| 137 | _cell_threshold.emplace_back(std::move(cell_threshold)); |
| 138 | _projection_threshold.emplace_back(std::move(projection_threshold)); |
| 139 | } |
| 140 | |
| 141 | protected: |
| 142 | LSTMLayerDataset() = default; |
| 143 | LSTMLayerDataset(LSTMLayerDataset &&) = default; |
| 144 | |
| 145 | private: |
| 146 | std::vector<TensorShape> _src_shapes{}; |
| 147 | std::vector<TensorShape> _input_weights_shapes{}; |
| 148 | std::vector<TensorShape> _recurrent_weights_shapes{}; |
| 149 | std::vector<TensorShape> _cell_bias_shapes{}; |
| 150 | std::vector<TensorShape> _output_cell_shapes{}; |
| 151 | std::vector<TensorShape> _dst_shapes{}; |
| 152 | std::vector<TensorShape> _scratch_shapes{}; |
| 153 | std::vector<ActivationLayerInfo> _infos{}; |
| 154 | std::vector<float> _cell_threshold{}; |
| 155 | std::vector<float> _projection_threshold{}; |
| 156 | }; |
| 157 | |
| 158 | class SmallLSTMLayerDataset final : public LSTMLayerDataset |
| 159 | { |
| 160 | public: |
| 161 | SmallLSTMLayerDataset() |
| 162 | { |
Georgios Pinitas | 3ada2b7 | 2018-08-23 15:54:36 +0100 | [diff] [blame] | 163 | add_config(TensorShape(8U), TensorShape(8U, 16U), TensorShape(16U, 16U), TensorShape(16U), TensorShape(16U), TensorShape(16U), TensorShape(64U), |
| 164 | ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU), 0.05f, 0.93f); |
| 165 | add_config(TensorShape(8U, 2U), TensorShape(8U, 16U), TensorShape(16U, 16U), TensorShape(16U), TensorShape(16U, 2U), TensorShape(16U, 2U), TensorShape(64U, 2U), |
| 166 | ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU), 0.05f, 0.93f); |
| 167 | add_config(TensorShape(8U, 2U), TensorShape(8U, 16U), TensorShape(16U, 16U), TensorShape(16U), TensorShape(16U, 2U), TensorShape(16U, 2U), TensorShape(48U, 2U), |
| 168 | ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU), 0.05f, 0.93f); |
Michalis Spyrou | bcedf51 | 2018-03-22 14:55:08 +0000 | [diff] [blame] | 169 | } |
| 170 | }; |
| 171 | |
| 172 | } // namespace datasets |
| 173 | } // namespace test |
| 174 | } // namespace arm_compute |
| 175 | #endif /* ARM_COMPUTE_TEST_LSTM_LAYER_DATASET */ |