1 /*
2 * Copyright (c) 2017-2021 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 #include "arm_compute/graph.h"
25 #include "support/ToolchainSupport.h"
26 #include "utils/CommonGraphOptions.h"
27 #include "utils/GraphUtils.h"
28 #include "utils/Utils.h"
29
30 using namespace arm_compute::utils;
31 using namespace arm_compute::graph::frontend;
32 using namespace arm_compute::graph_utils;
33
34 /** Example demonstrating how to implement LeNet's network using the Compute Library's graph API */
35 class GraphLenetExample : public Example
36 {
37 public:
GraphLenetExample()38 GraphLenetExample()
39 : cmd_parser(), common_opts(cmd_parser), common_params(), graph(0, "LeNet")
40 {
41 }
do_setup(int argc,char ** argv)42 bool do_setup(int argc, char **argv) override
43 {
44 // Parse arguments
45 cmd_parser.parse(argc, argv);
46 cmd_parser.validate();
47
48 // Consume common parameters
49 common_params = consume_common_graph_parameters(common_opts);
50
51 // Return when help menu is requested
52 if(common_params.help)
53 {
54 cmd_parser.print_help(argv[0]);
55 return false;
56 }
57
58 // Checks
59 ARM_COMPUTE_EXIT_ON_MSG(arm_compute::is_data_type_quantized_asymmetric(common_params.data_type), "QASYMM8 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;
66 unsigned int batches = 4; /** Number of batches */
67
68 // Create input descriptor
69 const auto operation_layout = common_params.data_layout;
70 const TensorShape tensor_shape = permute_shape(TensorShape(28U, 28U, 1U, batches), DataLayout::NCHW, operation_layout);
71 TensorDescriptor input_descriptor = TensorDescriptor(tensor_shape, common_params.data_type).set_layout(operation_layout);
72
73 // Set weights trained layout
74 const DataLayout weights_layout = DataLayout::NCHW;
75
76 //conv1 << pool1 << conv2 << pool2 << fc1 << act1 << fc2 << smx
77 graph << common_params.target
78 << common_params.fast_math_hint
79 << InputLayer(input_descriptor, get_input_accessor(common_params))
80 << ConvolutionLayer(
81 5U, 5U, 20U,
82 get_weights_accessor(data_path, "/cnn_data/lenet_model/conv1_w.npy", weights_layout),
83 get_weights_accessor(data_path, "/cnn_data/lenet_model/conv1_b.npy"),
84 PadStrideInfo(1, 1, 0, 0))
85 .set_name("conv1")
86 << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 2, operation_layout, PadStrideInfo(2, 2, 0, 0))).set_name("pool1")
87 << ConvolutionLayer(
88 5U, 5U, 50U,
89 get_weights_accessor(data_path, "/cnn_data/lenet_model/conv2_w.npy", weights_layout),
90 get_weights_accessor(data_path, "/cnn_data/lenet_model/conv2_b.npy"),
91 PadStrideInfo(1, 1, 0, 0))
92 .set_name("conv2")
93 << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 2, operation_layout, PadStrideInfo(2, 2, 0, 0))).set_name("pool2")
94 << FullyConnectedLayer(
95 500U,
96 get_weights_accessor(data_path, "/cnn_data/lenet_model/ip1_w.npy", weights_layout),
97 get_weights_accessor(data_path, "/cnn_data/lenet_model/ip1_b.npy"))
98 .set_name("ip1")
99 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU)).set_name("relu")
100 << FullyConnectedLayer(
101 10U,
102 get_weights_accessor(data_path, "/cnn_data/lenet_model/ip2_w.npy", weights_layout),
103 get_weights_accessor(data_path, "/cnn_data/lenet_model/ip2_b.npy"))
104 .set_name("ip2")
105 << SoftmaxLayer().set_name("prob")
106 << OutputLayer(get_output_accessor(common_params));
107
108 // Finalize graph
109 GraphConfig config;
110 config.num_threads = common_params.threads;
111 config.use_tuner = common_params.enable_tuner;
112 config.tuner_mode = common_params.tuner_mode;
113 config.tuner_file = common_params.tuner_file;
114 config.mlgo_file = common_params.mlgo_file;
115
116 graph.finalize(common_params.target, config);
117
118 return true;
119 }
do_run()120 void do_run() override
121 {
122 // Run graph
123 graph.run();
124 }
125
126 private:
127 CommandLineParser cmd_parser;
128 CommonGraphOptions common_opts;
129 CommonGraphParams common_params;
130 Stream graph;
131 };
132
133 /** Main program for LeNet
134 *
135 * Model is based on:
136 * http://yann.lecun.com/exdb/publis/pdf/lecun-98.pdf
137 * "Gradient-Based Learning Applied to Document Recognition"
138 * Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner
139 *
140 * The original model uses tanh instead of relu activations. However the use of relu activations in lenet has been
141 * widely adopted to improve accuracy.*
142 *
143 * @note To list all the possible arguments execute the binary appended with the --help option
144 *
145 * @param[in] argc Number of arguments
146 * @param[in] argv Arguments
147 */
main(int argc,char ** argv)148 int main(int argc, char **argv)
149 {
150 return arm_compute::utils::run_example<GraphLenetExample>(argc, argv);
151 }
152