Michalis Spyrou | 36a559e | 2018-03-20 10:30:58 +0000 | [diff] [blame] | 1 | /* |
| 2 | * Copyright (c) 2018 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_RNN_LAYER_FIXTURE |
| 25 | #define ARM_COMPUTE_TEST_RNN_LAYER_FIXTURE |
| 26 | |
| 27 | #include "tests/Globals.h" |
| 28 | #include "tests/framework/Asserts.h" |
| 29 | #include "tests/framework/Fixture.h" |
| 30 | #include "tests/validation/reference/ActivationLayer.h" |
| 31 | #include "tests/validation/reference/ArithmeticAddition.h" |
| 32 | #include "tests/validation/reference/FullyConnectedLayer.h" |
| 33 | #include "tests/validation/reference/GEMM.h" |
| 34 | |
| 35 | namespace arm_compute |
| 36 | { |
| 37 | namespace test |
| 38 | { |
| 39 | namespace validation |
| 40 | { |
| 41 | template <typename TensorType, typename AccessorType, typename FunctionType, typename T> |
| 42 | class RNNLayerValidationFixture : public framework::Fixture |
| 43 | { |
| 44 | public: |
| 45 | template <typename...> |
| 46 | void setup(TensorShape input_shape, TensorShape weights_shape, TensorShape recurrent_weights_shape, TensorShape bias_shape, TensorShape output_shape, ActivationLayerInfo info, |
| 47 | DataType data_type) |
| 48 | { |
| 49 | _target = compute_target(input_shape, weights_shape, recurrent_weights_shape, bias_shape, output_shape, info, data_type); |
| 50 | _reference = compute_reference(input_shape, weights_shape, recurrent_weights_shape, bias_shape, output_shape, info, data_type); |
| 51 | } |
| 52 | |
| 53 | protected: |
| 54 | template <typename U> |
| 55 | void fill(U &&tensor, int i) |
| 56 | { |
| 57 | std::uniform_real_distribution<> distribution(-1.0f, 1.0f); |
| 58 | library->fill(tensor, distribution, i); |
| 59 | } |
| 60 | |
| 61 | TensorType compute_target(const TensorShape &input_shape, const TensorShape &weights_shape, const TensorShape &recurrent_weights_shape, const TensorShape &bias_shape, const TensorShape &output_shape, |
| 62 | ActivationLayerInfo info, DataType data_type) |
| 63 | { |
| 64 | // Create tensors |
| 65 | TensorType input = create_tensor<TensorType>(input_shape, data_type); |
| 66 | TensorType weights = create_tensor<TensorType>(weights_shape, data_type); |
| 67 | TensorType recurrent_weights = create_tensor<TensorType>(recurrent_weights_shape, data_type); |
| 68 | TensorType bias = create_tensor<TensorType>(bias_shape, data_type); |
| 69 | TensorType hidden_state = create_tensor<TensorType>(output_shape, data_type); |
| 70 | TensorType output = create_tensor<TensorType>(output_shape, data_type); |
| 71 | |
| 72 | // Create and configure function |
| 73 | FunctionType rnn; |
| 74 | rnn.configure(&input, &weights, &recurrent_weights, &bias, &hidden_state, &output, info); |
| 75 | |
| 76 | ARM_COMPUTE_EXPECT(input.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 77 | ARM_COMPUTE_EXPECT(weights.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 78 | ARM_COMPUTE_EXPECT(recurrent_weights.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 79 | ARM_COMPUTE_EXPECT(bias.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 80 | ARM_COMPUTE_EXPECT(hidden_state.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 81 | ARM_COMPUTE_EXPECT(output.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 82 | |
| 83 | // Allocate tensors |
| 84 | input.allocator()->allocate(); |
| 85 | weights.allocator()->allocate(); |
| 86 | recurrent_weights.allocator()->allocate(); |
| 87 | bias.allocator()->allocate(); |
| 88 | hidden_state.allocator()->allocate(); |
| 89 | output.allocator()->allocate(); |
| 90 | |
| 91 | ARM_COMPUTE_EXPECT(!input.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 92 | ARM_COMPUTE_EXPECT(!weights.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 93 | ARM_COMPUTE_EXPECT(!recurrent_weights.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 94 | ARM_COMPUTE_EXPECT(!bias.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 95 | ARM_COMPUTE_EXPECT(!hidden_state.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 96 | ARM_COMPUTE_EXPECT(!output.info()->is_resizable(), framework::LogLevel::ERRORS); |
| 97 | |
| 98 | // Fill tensors |
| 99 | fill(AccessorType(input), 0); |
| 100 | fill(AccessorType(weights), 0); |
| 101 | fill(AccessorType(recurrent_weights), 0); |
| 102 | fill(AccessorType(bias), 0); |
| 103 | fill(AccessorType(hidden_state), 0); |
| 104 | |
| 105 | // Compute function |
| 106 | rnn.run(); |
| 107 | |
| 108 | return output; |
| 109 | } |
| 110 | |
| 111 | SimpleTensor<T> compute_reference(const TensorShape &input_shape, const TensorShape &weights_shape, const TensorShape &recurrent_weights_shape, const TensorShape &bias_shape, |
| 112 | const TensorShape &output_shape, ActivationLayerInfo info, DataType data_type) |
| 113 | { |
| 114 | // Create reference |
| 115 | SimpleTensor<T> input{ input_shape, data_type }; |
| 116 | SimpleTensor<T> weights{ weights_shape, data_type }; |
| 117 | SimpleTensor<T> recurrent_weights{ recurrent_weights_shape, data_type }; |
| 118 | SimpleTensor<T> bias{ bias_shape, data_type }; |
| 119 | SimpleTensor<T> hidden_state{ output_shape, data_type }; |
| 120 | |
| 121 | // Fill reference |
| 122 | fill(input, 0); |
| 123 | fill(weights, 0); |
| 124 | fill(recurrent_weights, 0); |
| 125 | fill(bias, 0); |
| 126 | fill(hidden_state, 0); |
| 127 | |
| 128 | TensorShape out_shape = recurrent_weights_shape; |
| 129 | out_shape.set(1, output_shape.y()); |
| 130 | |
| 131 | // Compute reference |
| 132 | SimpleTensor<T> out_w{ out_shape, data_type }; |
| 133 | SimpleTensor<T> fully_connected = reference::fully_connected_layer(input, weights, bias, out_shape); |
| 134 | SimpleTensor<T> gemm = reference::gemm(hidden_state, recurrent_weights, out_w, 1.f, 0.f); |
| 135 | SimpleTensor<T> add_res = reference::arithmetic_addition(fully_connected, gemm, data_type, ConvertPolicy::SATURATE); |
| 136 | return reference::activation_layer(add_res, info); |
| 137 | } |
| 138 | |
| 139 | TensorType _target{}; |
| 140 | SimpleTensor<T> _reference{}; |
| 141 | }; |
| 142 | } // namespace validation |
| 143 | } // namespace test |
| 144 | } // namespace arm_compute |
| 145 | #endif /* ARM_COMPUTE_TEST_RNN_LAYER_FIXTURE */ |