-
Notifications
You must be signed in to change notification settings - Fork 0
/
main.cpp
61 lines (45 loc) · 1.33 KB
/
main.cpp
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
#include "dense.hpp"
#include "dataset.hpp"
#include "activation.hpp"
#include "network.hpp"
#include "trainer.hpp"
#include "loss.hpp"
#include "metric.hpp"
#include "flatten.hpp"
#include "log.hpp"
#include "tensor.hpp"
#include "dropout.hpp"
#include "convolution.hpp"
#include "pool.hpp"
#include <iostream>
int main(__attribute__((unused)) int argc, __attribute__((unused)) char **argv)
{
bunji::Dataset dataset("scripts/dataset.json");
bunji::Network network;
bunji::Convolution c0(16, 5, 5, 1, 1);
bunji::Sigmoid a0;
bunji::MaxPool p0(2, 2);
bunji::Flatten f0;
bunji::Dense d0(256);
bunji::Sigmoid a2;
bunji::Dropout r0(0.35);
bunji::Dense d1(10);
bunji::Softmax a3;
network.add_layer(&c0);
network.add_layer(&a0);
network.add_layer(&p0);
network.add_layer(&f0);
network.add_layer(&d0);
network.add_layer(&a2);
network.add_layer(&r0);
network.add_layer(&d1);
network.add_layer(&a3);
network.build(std::make_tuple(28, 28, 1));
bunji::Crossentropy loss;
bunji::Crossentropy loss_metric;
bunji::Accuracy acc_metric;
std::vector<bunji::Metric*> metrics = {&loss_metric, &acc_metric};
bunji::Trainer network_trainer(&network, &dataset, &loss, metrics, 0.1);
network_trainer.fit(100, 32);
return 0;
}