cpp_neural_net_lib/test.cpp at master · casawa/cpp_neural_net_lib · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
#include "activations.h"
#include "graph.h"
#include "layers.h"
#include "losses.h"
#include "optimizers.h"
#include "xtensor/xarray.hpp"
#include "xtensor/xrandom.hpp"
#include "xtensor/xio.hpp"
#include <string>

/* Reports the results of testing a model on some test input. */
void graph_test_on_input(Graph &graph, xt::xarray<double> test) {
  std::cout << "Testing On: " << test << std::endl;
  std::cout << "Result: " << graph.run(test) << std::endl;
}

bool test_graph_optimization() {
  Graph graph;
  SGDOptimizer sgd_fc1(0.00003), sgd_fc2(0.00003), sgd_fc3(0.00003);
  graph.add_layer(new FullyConnected(3, 4, &sgd_fc1));
  graph.add_layer(new FullyConnected(4, 4, &sgd_fc2));
  graph.add_layer(new FullyConnected(4, 3, &sgd_fc3));

  MSELoss mse(3);
  xt::xarray<double> inp = xt::random::randn<double>({100, 3});
  xt::xarray<double> target = 2 * inp;

  graph.optimize(&mse, inp, target, 400, true);

  xt::xarray<double> test1 {{1.0, -2.0, 3.0}};
  graph_test_on_input(graph, test1);

  xt::xarray<double> test2 {{3.0, 24.0, -5.0}};
  graph_test_on_input(graph, test2);

  return true;
}

/* TODO: use expected arrays for the results */
bool test_relu() {
  /* Test ReLU forwards. */
  ReLU relu;
  xt::xarray<double> inp {1.0, -2.0, 3.0, 0.0};
  xt::xarray<double> fwd_result = relu.forward(inp);

  std::cout << "FORWARDS (input, result)" << std::endl;
  std::cout << inp << std::endl;
  std::cout << fwd_result << std::endl;

  /* Test ReLU backwards. */
  xt::xarray<double> inc_grad {3.0, 0.5, -2.0, 1.0};
  xt::xarray<double> bwd_result = relu.backward(inc_grad);

  std::cout << "BACKWARDS (inc grad, result)" << std::endl;
  std::cout << inc_grad << std::endl;
  std::cout << bwd_result << std::endl;

  return true;
}

/* TODO: use expected arrays for the results */
bool test_mse_loss() {
  /* Test MSE loss. */
  MSELoss mse(3);

  xt::xarray<double> inp {{1.0, -3.0, 3.0}};
  xt::xarray<double> targets {{2.0, -1.0, 3.0}};
  xt::xarray<double> fwd_result = mse.forward(inp, targets);

  std::cout << "FORWARDS (input, targets, result)" << std::endl;
  std::cout << inp << std::endl;
  std::cout << targets << std::endl;
  std::cout << fwd_result << std::endl;

  /* Test FC backwards. */
  xt::xarray<double> bwd_result = mse.backward();

  std::cout << "BACKWARDS (result)" << std::endl;
  std::cout << bwd_result << std::endl;

  return true;
}

/* TODO: use expected arrays for the results */
bool test_fc() {
  /* Test FC forwards. */
  SGDOptimizer sgd(0.03);
  FullyConnected fc(3, 5, &sgd);

  xt::xarray<double> inp {{1.0, -2.0, 3.0}};
  xt::xarray<double> fwd_result = fc.forward(inp);

  std::cout << "FORWARDS (input, result)" << std::endl;
  std::cout << inp << std::endl;
  std::cout << fwd_result << std::endl;

  /* Test FC backwards. */
  xt::xarray<double> inc_grad {{3.0, 0.5, -2.0, 1.0, 0.0}};
  xt::xarray<double> bwd_result = fc.backward(inc_grad);

  std::cout << "BACKWARDS (inc grad, result)" << std::endl;
  std::cout << inc_grad << std::endl;
  std::cout << bwd_result << std::endl;

  return true;
}


/* TODO: use expected arrays for the results */
bool test_sigmoid() {
  /* Test Sigmoid forwards. */
  Sigmoid sig;
  xt::xarray<double> inp {1.0, -2.0, 3.0, 0.0};
  xt::xarray<double> fwd_result = sig.forward(inp);

  std::cout << "FORWARDS (input, result)" << std::endl;
  std::cout << inp << std::endl;
  std::cout << fwd_result << std::endl;

  /* Test Sigmoid backwards. */
  xt::xarray<double> inc_grad {3.0, 0.5, -2.0, 1.0};
  xt::xarray<double> bwd_result = sig.backward(inc_grad);

  std::cout << "BACKWARDS (inc grad, result)" << std::endl;
  std::cout << inc_grad << std::endl;
  std::cout << bwd_result << std::endl;

  return true;
}

/* Runs a given test and reports the results. */
void run_test(bool (*test)(void), std::string test_name) {
  if (test()) {
    std::cout << test_name << " Tests PASSED" << std::endl;
  } else {
    std::cout << test_name << " Tests FAILED" << std::endl;
  }
}

int main() {
  run_test(test_relu, "ReLU");
  run_test(test_sigmoid, "Sigmoid");
  run_test(test_fc, "Fully Connected");
  run_test(test_mse_loss, "MSE Loss");
  run_test(test_graph_optimization, "Graph Optimization");
  return 0;
}