Gitiles
Code Review Sign In
gerrit-public.fairphone.software / platform / external / ComputeLibrary / 3d2d44ef55ab6b08afda8be48301ce3c55c7bc67 / . / examples / graph_inception_v3.cpp
blob: d9b7b05c9a96cac082e6bb2072a2ad66acc4db44 [file] [log] [blame]
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]1/*
2 * Copyright (c) 2017-2018 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 */
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]24#include "arm_compute/graph.h"
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]25#include "support/ToolchainSupport.h"
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]26#include "utils/CommonGraphOptions.h"
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]27#include "utils/GraphUtils.h"
28#include "utils/Utils.h"
29
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]30using namespace arm_compute::utils;
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]31using namespace arm_compute::graph::frontend;
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]32using namespace arm_compute::graph_utils;
33
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]34/** Example demonstrating how to implement InceptionV3's network using the Compute Library's graph API */
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]35class InceptionV3Example : public Example
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]36{
37public:
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]38 InceptionV3Example()
39 : cmd_parser(), common_opts(cmd_parser), common_params(), graph(0, "InceptionV3")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]40 {
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]41 }
42 bool do_setup(int argc, char **argv) override
43 {
44 // Parse arguments
45 cmd_parser.parse(argc, argv);
46
47 // Consume common parameters
48 common_params = consume_common_graph_parameters(common_opts);
49
50 // Return when help menu is requested
51 if(common_params.help)
52 {
53 cmd_parser.print_help(argv[0]);
54 return false;
55 }
56
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]57 // Checks
58 ARM_COMPUTE_EXIT_ON_MSG(arm_compute::is_data_type_quantized_asymmetric(common_params.data_type), "QASYMM8 not supported for this graph");
59 ARM_COMPUTE_EXIT_ON_MSG(common_params.data_type == DataType::F16 && common_params.target == Target::NEON, "F16 NEON not supported for this graph");
60
61 // Print parameter values
62 std::cout << common_params << std::endl;
63
64 // Get trainable parameters data path
65 std::string data_path = common_params.data_path;
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]66
Anthony Barbier06ea0482018-02-22 15:45:35 +0000[diff] [blame]67 // Create a preprocessor object
68 std::unique_ptr<IPreprocessor> preprocessor = arm_compute::support::cpp14::make_unique<TFPreproccessor>();
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]69
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]70 // Create input descriptor
71 const TensorShape tensor_shape = permute_shape(TensorShape(299U, 299U, 3U, 1U), DataLayout::NCHW, common_params.data_layout);
72 TensorDescriptor input_descriptor = TensorDescriptor(tensor_shape, common_params.data_type).set_layout(common_params.data_layout);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]73
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]74 // Set weights trained layout
75 const DataLayout weights_layout = DataLayout::NCHW;
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]76
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]77 graph << common_params.target
78 << common_params.fast_math_hint
79 << InputLayer(input_descriptor, get_input_accessor(common_params, std::move(preprocessor), false))
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]80 << ConvolutionLayer(3U, 3U, 32U,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]81 get_weights_accessor(data_path, "/cnn_data/inceptionv3_model/Conv2d_1a_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]82 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr), PadStrideInfo(2, 2, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]83 .set_name("Conv2d_1a_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]84 << BatchNormalizationLayer(get_weights_accessor(data_path,
85 "/cnn_data/inceptionv3_model/Conv2d_1a_3x3_BatchNorm_moving_mean.npy"),
86 get_weights_accessor(data_path,
87 "/cnn_data/inceptionv3_model/Conv2d_1a_3x3_BatchNorm_moving_variance.npy"),
88 get_random_accessor(1.f, 1.f), get_weights_accessor(data_path,
89 "/cnn_data/inceptionv3_model/Conv2d_1a_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]90 0.001f)
91 .set_name("Conv2d_1a_3x3/BatchNorm/batchnorm")
92 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name("Conv2d_1a_3x3/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]93 << ConvolutionLayer(3U, 3U, 32U,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]94 get_weights_accessor(data_path, "/cnn_data/inceptionv3_model/Conv2d_2a_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]95 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr), PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]96 .set_name("Conv2d_2a_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]97 << BatchNormalizationLayer(get_weights_accessor(data_path,
98 "/cnn_data/inceptionv3_model/Conv2d_2a_3x3_BatchNorm_moving_mean.npy"),
99 get_weights_accessor(data_path,
100 "/cnn_data/inceptionv3_model/Conv2d_2a_3x3_BatchNorm_moving_variance.npy"),
101 get_random_accessor(1.f, 1.f), get_weights_accessor(data_path,
102 "/cnn_data/inceptionv3_model/Conv2d_2a_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]103 0.001f)
104 .set_name("Conv2d_2a_3x3/BatchNorm/batchnorm")
105 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name("Conv2d_2a_3x3/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]106
107 << ConvolutionLayer(3U, 3U, 64U,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]108 get_weights_accessor(data_path, "/cnn_data/inceptionv3_model/Conv2d_2b_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]109 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr), PadStrideInfo(1, 1, 1, 1))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]110 .set_name("Conv2d_2b_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]111 << BatchNormalizationLayer(get_weights_accessor(data_path,
112 "/cnn_data/inceptionv3_model/Conv2d_2b_3x3_BatchNorm_moving_mean.npy"),
113 get_weights_accessor(data_path,
114 "/cnn_data/inceptionv3_model/Conv2d_2b_3x3_BatchNorm_moving_variance.npy"),
115 get_random_accessor(1.f, 1.f), get_weights_accessor(data_path,
116 "/cnn_data/inceptionv3_model/Conv2d_2b_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]117 0.001f)
118 .set_name("Conv2d_2b_3x3/BatchNorm/batchnorm")
119 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name("Conv2d_2b_3x3/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]120
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]121 << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 3, PadStrideInfo(2, 2, 0, 0, DimensionRoundingType::CEIL))).set_name("MaxPool_3a_3x3/MaxPool")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]122
123 << ConvolutionLayer(1U, 1U, 80U,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]124 get_weights_accessor(data_path, "/cnn_data/inceptionv3_model/Conv2d_3b_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]125 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr), PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]126 .set_name("Conv2d_3b_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]127 << BatchNormalizationLayer(get_weights_accessor(data_path,
128 "/cnn_data/inceptionv3_model/Conv2d_3b_1x1_BatchNorm_moving_mean.npy"),
129 get_weights_accessor(data_path,
130 "/cnn_data/inceptionv3_model/Conv2d_3b_1x1_BatchNorm_moving_variance.npy"),
131 get_random_accessor(1.f, 1.f), get_weights_accessor(data_path,
132 "/cnn_data/inceptionv3_model/Conv2d_3b_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]133 0.001f)
134 .set_name("Conv2d_3b_1x1/BatchNorm/batchnorm")
135 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name("Conv2d_3b_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]136
137 << ConvolutionLayer(3U, 3U, 192U,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]138 get_weights_accessor(data_path, "/cnn_data/inceptionv3_model/Conv2d_4a_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]139 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr), PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]140 .set_name("Conv2d_4a_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]141 << BatchNormalizationLayer(get_weights_accessor(data_path,
142 "/cnn_data/inceptionv3_model/Conv2d_4a_3x3_BatchNorm_moving_mean.npy"),
143 get_weights_accessor(data_path,
144 "/cnn_data/inceptionv3_model/Conv2d_4a_3x3_BatchNorm_moving_variance.npy"),
145 get_random_accessor(1.f, 1.f), get_weights_accessor(data_path,
146 "/cnn_data/inceptionv3_model/Conv2d_4a_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]147 0.001f)
148 .set_name("Conv2d_4a_3x3/BatchNorm/batchnorm")
149 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name("Conv2d_4a_3x3/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]150
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]151 << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 3, PadStrideInfo(2, 2, 0, 0, DimensionRoundingType::CEIL))).set_name("MaxPool_5a_3x3/MaxPool");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]152
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]153 graph << get_inception_node_A(data_path, "Mixed_5b", weights_layout, 64U, std::make_tuple(48U, 64U), std::make_tuple(64U, 96U, 96U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]154 32U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]155 .set_name("Mixed_5b/concat");
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]156 graph << get_inception_node_A(data_path, "Mixed_5c", weights_layout, 64U, std::make_tuple(48U, 64U), std::make_tuple(64U, 96U, 96U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]157 64U, true)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]158 .set_name("Mixed_5c/concat");
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]159 graph << get_inception_node_A(data_path, "Mixed_5d", weights_layout, 64U, std::make_tuple(48U, 64U), std::make_tuple(64U, 96U, 96U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]160 64U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]161 .set_name("Mixed_5d/concat");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]162
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]163 graph << get_inception_node_B(data_path, "Mixed_6a", weights_layout, 384U, std::make_tuple(64U, 96U, 96U)).set_name("Mixed_6a/concat");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]164
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]165 graph << get_inception_node_C(data_path, "Mixed_6b", weights_layout, 192U, std::make_tuple(128U, 128U, 192U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]166 std::make_tuple(128U, 128U, 128U, 128U, 192U), 192U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]167 .set_name("Mixed_6b/concat");
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]168 graph << get_inception_node_C(data_path, "Mixed_6c", weights_layout, 192U, std::make_tuple(160U, 160U, 192U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]169 std::make_tuple(160U, 160U, 160U, 160U, 192U), 192U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]170 .set_name("Mixed_6c/concat");
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]171 graph << get_inception_node_C(data_path, "Mixed_6d", weights_layout, 192U, std::make_tuple(160U, 160U, 192U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]172 std::make_tuple(160U, 160U, 160U, 160U, 192U), 192U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]173 .set_name("Mixed_6d/concat");
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]174 graph << get_inception_node_C(data_path, "Mixed_6e", weights_layout, 192U, std::make_tuple(192U, 192U, 192U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]175 std::make_tuple(192U, 192U, 192U, 192U, 192U), 192U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]176 .set_name("Mixed_6e/concat");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]177
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]178 graph << get_inception_node_D(data_path, "Mixed_7a", weights_layout, std::make_tuple(192U, 320U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]179 std::make_tuple(192U, 192U, 192U, 192U))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]180 .set_name("Mixed_7a/concat");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]181
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]182 graph << get_inception_node_E(data_path, "Mixed_7b", weights_layout, 320U, std::make_tuple(384U, 384U, 384U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]183 std::make_tuple(448U, 384U, 384U, 384U), 192U)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]184 .set_name("Mixed_7b/concat");
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]185 graph << get_inception_node_E(data_path, "Mixed_7c", weights_layout, 320U, std::make_tuple(384U, 384U, 384U),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]186 std::make_tuple(448U, 384U, 384U, 384U), 192U, true)
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]187 .set_name("Mixed_7c/concat");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]188
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]189 graph << PoolingLayer(PoolingLayerInfo(PoolingType::AVG, 8, PadStrideInfo(1, 1, 0, 0, DimensionRoundingType::CEIL))).set_name("Logits/AvgPool_1a_8x8/AvgPool")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]190 << ConvolutionLayer(1U, 1U, 1001U, get_weights_accessor(data_path,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]191 "/cnn_data/inceptionv3_model/Logits_Conv2d_1c_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]192 get_weights_accessor(data_path,
193 "/cnn_data/inceptionv3_model/Logits_Conv2d_1c_1x1_biases.npy"),
194 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]195 .set_name("Logits/Conv2d_1c_1x1/convolution")
196 << ReshapeLayer(TensorShape(1001U)).set_name("Predictions/Reshape")
197 << SoftmaxLayer().set_name("Predictions/Softmax")
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]198 << OutputLayer(get_output_accessor(common_params, 5));
Anthony Barbier06ea0482018-02-22 15:45:35 +0000[diff] [blame]199
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]200 // Finalize graph
201 GraphConfig config;
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]202 config.num_threads = common_params.threads;
203 config.use_tuner = common_params.enable_tuner;
204 config.tuner_file = common_params.tuner_file;
205
206 graph.finalize(common_params.target, config);
207
208 return true;
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]209 }
210
211 void do_run() override
212 {
213 graph.run();
214 }
215
216private:
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]217 CommandLineParser cmd_parser;
218 CommonGraphOptions common_opts;
219 CommonGraphParams common_params;
220 Stream graph;
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]221
222private:
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]223 ConcatLayer get_inception_node_A(const std::string &data_path, std::string &&param_path, DataLayout weights_layout,
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]224 unsigned int a_filt,
225 std::tuple<unsigned int, unsigned int> b_filters,
226 std::tuple<unsigned int, unsigned int, unsigned int> c_filters,
227 unsigned int d_filt,
228 bool is_name_different = false)
229 {
230 std::string total_path = "/cnn_data/inceptionv3_model/" + param_path + "_";
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]231
232 // This is due to a naming issue in the tf model
233 std::string conv_id0 = "_0a_";
234 std::string conv_id1 = "2d_0b_";
235 if(is_name_different)
236 {
237 conv_id0 = "_0b_";
238 conv_id1 = "_1_0c_";
239 }
240
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]241 SubStream i_a(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]242 i_a << ConvolutionLayer(
243 1U, 1U, a_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]244 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]245 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
246 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]247 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]248 << BatchNormalizationLayer(
249 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
250 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
251 get_random_accessor(1.f, 1.f),
252 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]253 0.001f)
254 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/BatchNorm/batchnorm")
255 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_0a_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]256
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]257 SubStream i_b(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]258 i_b << ConvolutionLayer(
259 1U, 1U, std::get<0>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]260 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id0 + "1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]261 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
262 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]263 .set_name(param_path + "/Branch_1/Conv2d" + conv_id0 + "1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]264 << BatchNormalizationLayer(
265 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id0 + "1x1_BatchNorm_moving_mean.npy"),
266 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id0 + "1x1_BatchNorm_moving_variance.npy"),
267 get_random_accessor(1.f, 1.f),
268 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id0 + "1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]269 0.001f)
270 .set_name(param_path + "/Branch_1/Conv2d" + conv_id0 + "1x1/BatchNorm/batchnorm")
271 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d" + conv_id0 + "1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]272 << ConvolutionLayer(
273 5U, 5U, std::get<1>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]274 get_weights_accessor(data_path, total_path + "Branch_1_Conv" + conv_id1 + "5x5_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]275 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
276 PadStrideInfo(1, 1, 2, 2))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]277 .set_name(param_path + "/Branch_1/Conv2d" + conv_id1 + "5x5/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]278 << BatchNormalizationLayer(
279 get_weights_accessor(data_path, total_path + "Branch_1_Conv" + conv_id1 + "5x5_BatchNorm_moving_mean.npy"),
280 get_weights_accessor(data_path, total_path + "Branch_1_Conv" + conv_id1 + "5x5_BatchNorm_moving_variance.npy"),
281 get_random_accessor(1.f, 1.f),
282 get_weights_accessor(data_path, total_path + "Branch_1_Conv" + conv_id1 + "5x5_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]283 0.001f)
284 .set_name(param_path + "/Branch_1/Conv2d" + conv_id1 + "5x5/BatchNorm/batchnorm")
285 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d" + conv_id1 + "5x5/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]286
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]287 SubStream i_c(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]288 i_c << ConvolutionLayer(
289 1U, 1U, std::get<0>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]290 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]291 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
292 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]293 .set_name(param_path + "/Branch_2/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]294 << BatchNormalizationLayer(
295 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
296 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
297 get_random_accessor(1.f, 1.f),
298 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]299 0.001f)
300 .set_name(param_path + "/Branch_2/Conv2d_0a_1x1/BatchNorm/batchnorm")
301 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]302 << ConvolutionLayer(
303 3U, 3U, std::get<1>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]304 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]305 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
306 PadStrideInfo(1, 1, 1, 1))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]307 .set_name(param_path + "/Branch_2/Conv2d_0b_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]308 << BatchNormalizationLayer(
309 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_BatchNorm_moving_mean.npy"),
310 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_BatchNorm_moving_variance.npy"),
311 get_random_accessor(1.f, 1.f),
312 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]313 0.001f)
314 .set_name(param_path + "/Branch_2/Conv2d_0b_3x3/BatchNorm/batchnorm")
315 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0b_3x3/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]316 << ConvolutionLayer(
317 3U, 3U, std::get<2>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]318 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]319 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
320 PadStrideInfo(1, 1, 1, 1))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]321 .set_name(param_path + "/Branch_2/Conv2d_0c_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]322 << BatchNormalizationLayer(
323 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_3x3_BatchNorm_moving_mean.npy"),
324 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_3x3_BatchNorm_moving_variance.npy"),
325 get_random_accessor(1.f, 1.f),
326 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]327 0.001f)
328 .set_name(param_path + "/Branch_2/Conv2d_0c_3x3/BatchNorm/batcnorm")
329 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0c_3x3/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]330
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]331 SubStream i_d(graph);
332 i_d << PoolingLayer(PoolingLayerInfo(PoolingType::AVG, 3, PadStrideInfo(1, 1, 1, 1, DimensionRoundingType::CEIL), true)).set_name(param_path + "/Branch_3/AvgPool_0a_3x3/AvgPool")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]333 << ConvolutionLayer(
334 1U, 1U, d_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]335 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]336 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
337 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]338 .set_name(param_path + "/Branch_3/Conv2d_0b_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]339 << BatchNormalizationLayer(
340 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_moving_mean.npy"),
341 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_moving_variance.npy"),
342 get_random_accessor(1.f, 1.f),
343 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]344 0.001f)
345 .set_name(param_path + "/Branch_3/Conv2d_0b_1x1/BatchNorm/batchnorm")
346 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_3/Conv2d_0b_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]347
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]348 return ConcatLayer(std::move(i_a), std::move(i_b), std::move(i_c), std::move(i_d));
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]349 }
350
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]351 ConcatLayer get_inception_node_B(const std::string &data_path, std::string &&param_path, DataLayout weights_layout,
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]352 unsigned int a_filt,
353 std::tuple<unsigned int, unsigned int, unsigned int> b_filters)
354 {
355 std::string total_path = "/cnn_data/inceptionv3_model/" + param_path + "_";
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]356 SubStream i_a(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]357 i_a << ConvolutionLayer(
358 3U, 3U, a_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]359 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]360 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
361 PadStrideInfo(2, 2, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]362 .set_name(param_path + "/Branch_0/Conv2d_1a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]363 << BatchNormalizationLayer(
364 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_1x1_BatchNorm_moving_mean.npy"),
365 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_1x1_BatchNorm_moving_variance.npy"),
366 get_random_accessor(1.f, 1.f),
367 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]368 0.001f)
369 .set_name(param_path + "/Branch_0/Conv2d_1a_1x1/BatchNorm/batchnorm")
370 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_1a_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]371
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]372 SubStream i_b(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]373 i_b << ConvolutionLayer(
374 1U, 1U, std::get<0>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]375 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]376 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
377 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]378 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]379 << BatchNormalizationLayer(
380 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
381 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
382 get_random_accessor(1.f, 1.f),
383 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]384 0.001f)
385 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/BatchNorm/batchnorm")
386 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]387 << ConvolutionLayer(
388 3U, 3U, std::get<1>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]389 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]390 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
391 PadStrideInfo(1, 1, 1, 1))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]392 .set_name(param_path + "/Branch_1/Conv2d_0b_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]393 << BatchNormalizationLayer(
394 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_3x3_BatchNorm_moving_mean.npy"),
395 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_3x3_BatchNorm_moving_variance.npy"),
396 get_random_accessor(1.f, 1.f),
397 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]398 0.001f)
399 .set_name(param_path + "/Branch_1/Conv2d_0b_3x3/BatchNorm/batchnorm")
400 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0b_3x3/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]401 << ConvolutionLayer(
402 3U, 3U, std::get<2>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]403 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]404 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
405 PadStrideInfo(2, 2, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]406 .set_name(param_path + "/Branch_1/Conv2d_1a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]407 << BatchNormalizationLayer(
408 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_1x1_BatchNorm_moving_mean.npy"),
409 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_1x1_BatchNorm_moving_variance.npy"),
410 get_random_accessor(1.f, 1.f),
411 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]412 0.001f)
413 .set_name(param_path + "/Branch_1/Conv2d_1a_1x1/BatchNorm/batchnorm")
414 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_1a_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]415
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]416 SubStream i_c(graph);
417 i_c << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 3, PadStrideInfo(2, 2, 0, 0, DimensionRoundingType::CEIL))).set_name(param_path + "/Branch_2/MaxPool_1a_3x3/MaxPool");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]418
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]419 return ConcatLayer(std::move(i_a), std::move(i_b), std::move(i_c));
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]420 }
421
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]422 ConcatLayer get_inception_node_C(const std::string &data_path, std::string &&param_path, DataLayout weights_layout,
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]423 unsigned int a_filt,
424 std::tuple<unsigned int, unsigned int, unsigned int> b_filters,
425 std::tuple<unsigned int, unsigned int, unsigned int, unsigned int, unsigned int> c_filters,
426 unsigned int d_filt)
427 {
428 std::string total_path = "/cnn_data/inceptionv3_model/" + param_path + "_";
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]429 SubStream i_a(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]430 i_a << ConvolutionLayer(
431 1U, 1U, a_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]432 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]433 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
434 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]435 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]436 << BatchNormalizationLayer(
437 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
438 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
439 get_random_accessor(1.f, 1.f),
440 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]441 0.001f)
442 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/BatchNorm/batchnorm")
443 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_0a_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]444
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]445 SubStream i_b(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]446 i_b << ConvolutionLayer(
447 1U, 1U, std::get<0>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]448 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]449 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
450 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]451 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]452 << BatchNormalizationLayer(
453 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
454 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
455 get_random_accessor(1.f, 1.f),
456 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]457 0.001f)
458 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/BatchNorm/batchnorm")
459 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]460 << ConvolutionLayer(
461 7U, 1U, std::get<1>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]462 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]463 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
464 PadStrideInfo(1, 1, 3, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]465 .set_name(param_path + "/Branch_1/Conv2d_0b_1x7/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]466 << BatchNormalizationLayer(
467 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_BatchNorm_moving_mean.npy"),
468 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_BatchNorm_moving_variance.npy"),
469 get_random_accessor(1.f, 1.f),
470 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]471 0.001f)
472 .set_name(param_path + "/Branch_1/Conv2d_0b_1x7/BatchNorm/batchnorm")
473 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0b_1x7/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]474 << ConvolutionLayer(
475 1U, 7U, std::get<2>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]476 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]477 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
478 PadStrideInfo(1, 1, 0, 3))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]479 .set_name(param_path + "/Branch_1/Conv2d_0c_7x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]480 << BatchNormalizationLayer(
481 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_BatchNorm_moving_mean.npy"),
482 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_BatchNorm_moving_variance.npy"),
483 get_random_accessor(1.f, 1.f),
484 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]485 0.001f)
486 .set_name(param_path + "/Branch_1/Conv2d_0c_7x1/BatchNorm/batchnorm")
487 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_0c_7x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]488
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]489 SubStream i_c(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]490 i_c << ConvolutionLayer(
491 1U, 1U, std::get<0>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]492 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]493 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
494 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]495 .set_name(param_path + "/Branch_2/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]496 << BatchNormalizationLayer(
497 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
498 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
499 get_random_accessor(1.f, 1.f),
500 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]501 0.001f)
502 .set_name(param_path + "/Branch_2/Conv2d_0a_1x1/BatchNorm/batchnorm")
503 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]504 << ConvolutionLayer(
505 1U, 7U, std::get<1>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]506 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_7x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]507 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
508 PadStrideInfo(1, 1, 0, 3))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]509 .set_name(param_path + "/Branch_2/Conv2d_0b_7x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]510 << BatchNormalizationLayer(
511 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_7x1_BatchNorm_moving_mean.npy"),
512 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_7x1_BatchNorm_moving_variance.npy"),
513 get_random_accessor(1.f, 1.f),
514 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_7x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]515 0.001f)
516 .set_name(param_path + "/Branch_2/Conv2d_0b_7x1/BatchNorm/batchnorm")
517 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0b_7x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]518 << ConvolutionLayer(
519 7U, 1U, std::get<2>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]520 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x7_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]521 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
522 PadStrideInfo(1, 1, 3, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]523 .set_name(param_path + "/Branch_2/Conv2d_0c_1x7/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]524 << BatchNormalizationLayer(
525 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x7_BatchNorm_moving_mean.npy"),
526 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x7_BatchNorm_moving_variance.npy"),
527 get_random_accessor(1.f, 1.f),
528 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x7_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]529 0.001f)
530 .set_name(param_path + "/Branch_2/Conv2d_0c_1x7/BatchNorm/batchnorm")
531 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0c_1x7/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]532 << ConvolutionLayer(
533 1U, 7U, std::get<3>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]534 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_7x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]535 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
536 PadStrideInfo(1, 1, 0, 3))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]537 .set_name(param_path + "/Branch_2/Conv2d_0d_7x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]538 << BatchNormalizationLayer(
539 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_7x1_BatchNorm_moving_mean.npy"),
540 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_7x1_BatchNorm_moving_variance.npy"),
541 get_random_accessor(1.f, 1.f),
542 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_7x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]543 0.001f)
544 .set_name(param_path + "/Branch_2/Conv2d_0d_7x1/BatchNorm/batchnorm")
545 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0d_7x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]546 << ConvolutionLayer(
547 7U, 1U, std::get<4>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]548 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0e_1x7_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]549 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
550 PadStrideInfo(1, 1, 3, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]551 .set_name(param_path + "/Branch_2/Conv2d_0e_1x7/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]552 << BatchNormalizationLayer(
553 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0e_1x7_BatchNorm_moving_mean.npy"),
554 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0e_1x7_BatchNorm_moving_variance.npy"),
555 get_random_accessor(1.f, 1.f),
556 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0e_1x7_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]557 0.001f)
558 .set_name(param_path + "/Branch_2/Conv2d_0e_1x7/BatchNorm/batchnorm")
559 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0e_1x7/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]560
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]561 SubStream i_d(graph);
562 i_d << PoolingLayer(PoolingLayerInfo(PoolingType::AVG, 3, PadStrideInfo(1, 1, 1, 1, DimensionRoundingType::CEIL), true)).set_name(param_path + "/Branch_3/AvgPool_0a_3x3/AvgPool")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]563 << ConvolutionLayer(
564 1U, 1U, d_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]565 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]566 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
567 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]568 .set_name(param_path + "/Branch_3/Conv2d_0b_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]569 << BatchNormalizationLayer(
570 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_moving_mean.npy"),
571 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_moving_variance.npy"),
572 get_random_accessor(1.f, 1.f),
573 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]574 0.001f)
575 .set_name(param_path + "/Branch_3/Conv2d_0b_1x1/BatchNorm/batchnorm")
576 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_3/Conv2d_0b_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]577
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]578 return ConcatLayer(std::move(i_a), std::move(i_b), std::move(i_c), std::move(i_d));
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]579 }
580
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]581 ConcatLayer get_inception_node_D(const std::string &data_path, std::string &&param_path, DataLayout weights_layout,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]582 std::tuple<unsigned int, unsigned int> a_filters,
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]583 std::tuple<unsigned int, unsigned int, unsigned int, unsigned int> b_filters)
584 {
585 std::string total_path = "/cnn_data/inceptionv3_model/" + param_path + "_";
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]586 SubStream i_a(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]587 i_a << ConvolutionLayer(
588 1U, 1U, std::get<0>(a_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]589 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]590 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
591 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]592 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]593 << BatchNormalizationLayer(
594 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
595 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
596 get_random_accessor(1.f, 1.f),
597 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]598 0.001f)
599 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/BatchNorm/batchnorm")
600 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]601 << ConvolutionLayer(
602 3U, 3U, std::get<1>(a_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]603 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]604 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
605 PadStrideInfo(2, 2, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]606 .set_name(param_path + "/Branch_0/Conv2d_1a_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]607 << BatchNormalizationLayer(
608 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_3x3_BatchNorm_moving_mean.npy"),
609 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_3x3_BatchNorm_moving_variance.npy"),
610 get_random_accessor(1.f, 1.f),
611 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_1a_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]612 0.001f)
613 .set_name(param_path + "/Branch_0/Conv2d_1a_3x3/BatchNorm/batchnorm")
614 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_1a_3x3/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]615
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]616 SubStream i_b(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]617 i_b << ConvolutionLayer(
618 1U, 1U, std::get<0>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]619 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]620 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
621 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]622 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]623 << BatchNormalizationLayer(
624 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
625 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
626 get_random_accessor(1.f, 1.f),
627 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]628 0.001f)
629 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/BatchNorm/batchnorm")
630 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]631 << ConvolutionLayer(
632 7U, 1U, std::get<1>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]633 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]634 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
635 PadStrideInfo(1, 1, 3, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]636 .set_name(param_path + "/Branch_1/Conv2d_0b_1x7/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]637 << BatchNormalizationLayer(
638 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_BatchNorm_moving_mean.npy"),
639 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_BatchNorm_moving_variance.npy"),
640 get_random_accessor(1.f, 1.f),
641 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x7_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]642 0.001f)
643 .set_name(param_path + "/Branch_1/Conv2d_0b_1x7/BatchNorm/batchnorm")
644 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0b_1x7/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]645 << ConvolutionLayer(
646 1U, 7U, std::get<2>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]647 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]648 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
649 PadStrideInfo(1, 1, 0, 3))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]650 .set_name(param_path + "/Branch_1/Conv2d_0c_7x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]651 << BatchNormalizationLayer(
652 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_BatchNorm_moving_mean.npy"),
653 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_BatchNorm_moving_variance.npy"),
654 get_random_accessor(1.f, 1.f),
655 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0c_7x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]656 0.001f)
657 .set_name(param_path + "/Branch_1/Conv2d_0c_7x1/BatchNorm/batchnorm")
658 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0c_7x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]659 << ConvolutionLayer(
660 3U, 3U, std::get<3>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]661 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]662 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
663 PadStrideInfo(2, 2, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]664 .set_name(param_path + "/Branch_1/Conv2d_1a_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]665 << BatchNormalizationLayer(
666 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_3x3_BatchNorm_moving_mean.npy"),
667 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_3x3_BatchNorm_moving_variance.npy"),
668 get_random_accessor(1.f, 1.f),
669 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_1a_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]670 0.001f)
671 .set_name(param_path + "/Branch_1/Conv2d_1a_3x3/BatchNorm/batchnorm")
672 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_1a_3x3/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]673
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]674 SubStream i_c(graph);
675 i_c << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 3, PadStrideInfo(2, 2, 0, 0, DimensionRoundingType::CEIL))).set_name(param_path + "/Branch_2/MaxPool_1a_3x3/MaxPool");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]676
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]677 return ConcatLayer(std::move(i_a), std::move(i_b), std::move(i_c));
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]678 }
679
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]680 ConcatLayer get_inception_node_E(const std::string &data_path, std::string &&param_path, DataLayout weights_layout,
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]681 unsigned int a_filt,
682 std::tuple<unsigned int, unsigned int, unsigned int> b_filters,
683 std::tuple<unsigned int, unsigned int, unsigned int, unsigned int> c_filters,
684 unsigned int d_filt,
685 bool is_name_different = false)
686 {
687 // This is due to a naming issue in the tf model
688 std::string conv_id = "_0b_";
689 if(is_name_different)
690 {
691 conv_id = "_0c_";
692 }
693
694 std::string total_path = "/cnn_data/inceptionv3_model/" + param_path + "_";
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]695 SubStream i_a(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]696 i_a << ConvolutionLayer(
697 1U, 1U, a_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]698 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]699 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
700 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]701 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]702 << BatchNormalizationLayer(
703 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
704 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
705 get_random_accessor(1.f, 1.f),
706 get_weights_accessor(data_path, total_path + "Branch_0_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]707 0.001f)
708 .set_name(param_path + "/Branch_0/Conv2d_0a_1x1/BatchNorm/batchnorm")
709 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_0/Conv2d_0a_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]710
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]711 SubStream i_b(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]712 i_b << ConvolutionLayer(
713 1U, 1U, std::get<0>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]714 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]715 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
716 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]717 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]718 << BatchNormalizationLayer(
719 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
720 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
721 get_random_accessor(1.f, 1.f),
722 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]723 0.001f)
724 .set_name(param_path + "/Branch_1/Conv2d_0a_1x1/BatchNorm/batchnorm")
725 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0a_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]726
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]727 SubStream i_b1(i_b);
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]728 i_b1 << ConvolutionLayer(
729 3U, 1U, std::get<1>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]730 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]731 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
732 PadStrideInfo(1, 1, 1, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]733 .set_name(param_path + "/Branch_1/Conv2d_0b_1x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]734 << BatchNormalizationLayer(
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]735 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x3_BatchNorm_moving_mean.npy"),
736 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x3_BatchNorm_moving_variance.npy"),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]737 get_random_accessor(1.f, 1.f),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]738 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d_0b_1x3_BatchNorm_beta.npy"),
739 0.001f)
740 .set_name(param_path + "/Branch_1/Conv2d_0b_1x3/BatchNorm/batchnorm")
741 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d_0b_1x3/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]742
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]743 SubStream i_b2(i_b);
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]744 i_b2 << ConvolutionLayer(
745 1U, 3U, std::get<2>(b_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]746 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id + "3x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]747 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
748 PadStrideInfo(1, 1, 0, 1))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]749 .set_name(param_path + "/Branch_1/Conv2d" + conv_id + "3x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]750 << BatchNormalizationLayer(
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]751 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id + "3x1_BatchNorm_moving_mean.npy"),
752 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id + "3x1_BatchNorm_moving_variance.npy"),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]753 get_random_accessor(1.f, 1.f),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]754 get_weights_accessor(data_path, total_path + "Branch_1_Conv2d" + conv_id + "3x1_BatchNorm_beta.npy"),
755 0.001f)
756 .set_name(param_path + "/Branch_1/Conv2d" + conv_id + "3x1/BatchNorm/batchnorm")
757 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_1/Conv2d" + conv_id + "3x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]758
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]759 // Merge b1 and b2
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]760 i_b << ConcatLayer(std::move(i_b1), std::move(i_b2)).set_name(param_path + "/Branch_1/concat");
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]761
762 SubStream i_c(graph);
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]763 i_c << ConvolutionLayer(
764 1U, 1U, std::get<0>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]765 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]766 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
767 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]768 .set_name(param_path + "/Branch_2/Conv2d_0a_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]769 << BatchNormalizationLayer(
770 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_moving_mean.npy"),
771 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_moving_variance.npy"),
772 get_random_accessor(1.f, 1.f),
773 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0a_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]774 0.001f)
775 .set_name(param_path + "/Branch_2/Conv2d_0a_1x1/BatchNorm/batchnorm")
776 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0a_1x1/Relu")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]777 << ConvolutionLayer(
778 3U, 3U, std::get<1>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]779 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]780 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
781 PadStrideInfo(1, 1, 1, 1))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]782 .set_name(param_path + "/Branch_2/Conv2d_0b_3x3/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]783 << BatchNormalizationLayer(
784 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_BatchNorm_moving_mean.npy"),
785 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_BatchNorm_moving_variance.npy"),
786 get_random_accessor(1.f, 1.f),
787 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0b_3x3_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]788 0.001f)
789 .set_name(param_path + "/Branch_2/Conv2d_0b_3x3/BatchNorm/batchnorm")
790 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0b_3x3/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]791
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]792 SubStream i_c1(i_c);
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]793 i_c1 << ConvolutionLayer(
794 3U, 1U, std::get<2>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]795 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x3_weights.npy", weights_layout),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]796 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
797 PadStrideInfo(1, 1, 1, 0))
798 .set_name(param_path + "/Branch_2/Conv2d_0c_1x3/convolution")
799 << BatchNormalizationLayer(
800 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x3_BatchNorm_moving_mean.npy"),
801 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x3_BatchNorm_moving_variance.npy"),
802 get_random_accessor(1.f, 1.f),
803 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0c_1x3_BatchNorm_beta.npy"),
804 0.001f)
805 .set_name(param_path + "/Branch_2/Conv2d_0c_1x3/BatchNorm/batchnorm")
806 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0c_1x3/Relu");
807
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]808 SubStream i_c2(i_c);
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]809 i_c2 << ConvolutionLayer(
810 1U, 3U, std::get<3>(c_filters),
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]811 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_3x1_weights.npy", weights_layout),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]812 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
813 PadStrideInfo(1, 1, 0, 1))
814 .set_name(param_path + "/Branch_2/Conv2d_0d_3x1/convolution")
815 << BatchNormalizationLayer(
816 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_3x1_BatchNorm_moving_mean.npy"),
817 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_3x1_BatchNorm_moving_variance.npy"),
818 get_random_accessor(1.f, 1.f),
819 get_weights_accessor(data_path, total_path + "Branch_2_Conv2d_0d_3x1_BatchNorm_beta.npy"),
820 0.001f)
821 .set_name(param_path + "/Branch_2/Conv2d_0d_3x1/BatchNorm/batchnorm")
822 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_2/Conv2d_0d_3x1/Relu");
823
824 // Merge i_c1 and i_c2
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]825 i_c << ConcatLayer(std::move(i_c1), std::move(i_c2)).set_name(param_path + "/Branch_2/concat");
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]826
827 SubStream i_d(graph);
828 i_d << PoolingLayer(PoolingLayerInfo(PoolingType::AVG, 3, PadStrideInfo(1, 1, 1, 1, DimensionRoundingType::CEIL), true)).set_name(param_path + "/Branch_3/AvgPool_0a_3x3/AvgPool")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]829 << ConvolutionLayer(
830 1U, 1U, d_filt,
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]831 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_weights.npy", weights_layout),
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]832 std::unique_ptr<arm_compute::graph::ITensorAccessor>(nullptr),
833 PadStrideInfo(1, 1, 0, 0))
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]834 .set_name(param_path + "/Branch_3/Conv2d_0b_1x1/convolution")
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]835 << BatchNormalizationLayer(
836 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_moving_mean.npy"),
837 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_moving_variance.npy"),
838 get_random_accessor(1.f, 1.f),
839 get_weights_accessor(data_path, total_path + "Branch_3_Conv2d_0b_1x1_BatchNorm_beta.npy"),
Jenkinsb3a371b2018-05-23 11:36:53 +0100[diff] [blame]840 0.001f)
841 .set_name(param_path + "/Branch_3/Conv2d_0b_1x1/BatchNorm/batchnorm")
842 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name(param_path + "/Branch_3/Conv2d_0b_1x1/Relu");
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]843
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]844 return ConcatLayer(std::move(i_a), std::move(i_b), std::move(i_c), std::move(i_d));
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]845 }
846};
847
848/** Main program for Inception V3
849 *
Jenkinsb9abeae2018-11-22 11:58:08 +0000[diff] [blame]850 * Model is based on:
851 * https://arxiv.org/abs/1512.00567
852 * "Rethinking the Inception Architecture for Computer Vision"
853 * Christian Szegedy, Vincent Vanhoucke, Sergey Ioffe, Jonathon Shlens, Zbigniew Wojna
854 *
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]855 * @note To list all the possible arguments execute the binary appended with the --help option
856 *
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]857 * @param[in] argc Number of arguments
Jenkins52ba29e2018-08-29 15:32:11 +0000[diff] [blame]858 * @param[in] argv Arguments
Anthony Barbierf45d5a92018-01-24 16:23:15 +0000[diff] [blame]859 */
860int main(int argc, char **argv)
861{
862 return arm_compute::utils::run_example<InceptionV3Example>(argc, argv);
863}