2 * Copyright (c) 2017 ARM Limited.
4 * SPDX-License-Identifier: MIT
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:
13 * The above copyright notice and this permission notice shall be included in all
14 * copies or substantial portions of the Software.
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
24 #ifndef ARM_COMPUTE_CL /* Needed by Utils.cpp to handle OpenCL exceptions properly */
25 #error "This example needs to be built with -DARM_COMPUTE_CL"
26 #endif /* ARM_COMPUTE_CL */
28 #include "arm_compute/core/Logger.h"
29 #include "arm_compute/graph/Graph.h"
30 #include "arm_compute/graph/Nodes.h"
31 #include "arm_compute/runtime/CL/CLScheduler.h"
32 #include "arm_compute/runtime/Scheduler.h"
33 #include "support/ToolchainSupport.h"
34 #include "utils/GraphUtils.h"
35 #include "utils/Utils.h"
41 using namespace arm_compute::graph;
42 using namespace arm_compute::graph_utils;
44 /** Generates appropriate accessor according to the specified path
46 * @note If path is empty will generate a DummyAccessor else will generate a NumPyBinLoader
48 * @param path Path to the data files
49 * @param data_file Relative path to the data files from path
51 * @return An appropriate tensor accessor
53 std::unique_ptr<ITensorAccessor> get_accessor(const std::string &path, const std::string &data_file)
57 return arm_compute::support::cpp14::make_unique<DummyAccessor>();
61 return arm_compute::support::cpp14::make_unique<NumPyBinLoader>(path + data_file);
65 /** Example demonstrating how to implement LeNet's network using the Compute Library's graph API
67 * @param[in] argc Number of arguments
68 * @param[in] argv Arguments ( [optional] Path to the weights folder, [optional] batches )
70 void main_graph_lenet(int argc, const char **argv)
72 std::string data_path; /** Path to the trainable data */
73 unsigned int batches = 4; /** Number of batches */
79 std::cout << "Usage: " << argv[0] << " [path_to_data] [batches]\n\n";
80 std::cout << "No data folder provided: using random values\n\n";
84 //Do something with argv[1]
86 std::cout << "Usage: " << argv[0] << " [path_to_data] [batches]\n\n";
87 std::cout << "No number of batches where specified, thus will use the default : " << batches << "\n\n";
91 //Do something with argv[1] and argv[2]
93 batches = std::strtol(argv[2], nullptr, 0);
96 // Check if OpenCL is available and initialize the scheduler
97 TargetHint hint = TargetHint::NEON;
98 if(arm_compute::opencl_is_available())
100 arm_compute::CLScheduler::get().default_init();
101 hint = TargetHint::OPENCL;
105 arm_compute::Logger::get().set_logger(std::cout, arm_compute::LoggerVerbosity::INFO);
107 //conv1 << pool1 << conv2 << pool2 << fc1 << act1 << fc2 << smx
109 << Tensor(TensorInfo(TensorShape(28U, 28U, 1U, batches), 1, DataType::F32), DummyAccessor())
112 get_accessor(data_path, "/cnn_data/lenet_model/conv1_w.npy"),
113 get_accessor(data_path, "/cnn_data/lenet_model/conv1_b.npy"),
114 PadStrideInfo(1, 1, 0, 0))
115 << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 2, PadStrideInfo(2, 2, 0, 0)))
118 get_accessor(data_path, "/cnn_data/lenet_model/conv2_w.npy"),
119 get_accessor(data_path, "/cnn_data/lenet_model/conv2_b.npy"),
120 PadStrideInfo(1, 1, 0, 0))
121 << PoolingLayer(PoolingLayerInfo(PoolingType::MAX, 2, PadStrideInfo(2, 2, 0, 0)))
122 << FullyConnectedLayer(
124 get_accessor(data_path, "/cnn_data/lenet_model/ip1_w.npy"),
125 get_accessor(data_path, "/cnn_data/lenet_model/ip1_b.npy"))
126 << ActivationLayer(ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU))
127 << FullyConnectedLayer(
129 get_accessor(data_path, "/cnn_data/lenet_model/ip2_w.npy"),
130 get_accessor(data_path, "/cnn_data/lenet_model/ip2_b.npy"))
132 << Tensor(DummyAccessor());
137 /** Main program for LeNet
139 * @param[in] argc Number of arguments
140 * @param[in] argv Arguments ( [optional] Path to the weights folder, [optional] batches )
142 int main(int argc, const char **argv)
144 return arm_compute::utils::run_example(argc, argv, main_graph_lenet);