tiny_dnn 1.0.0
A header only, dependency-free deep learning framework in C++11
Loading...
Searching...
No Matches
concat_layer.h
1/*
2 Copyright (c) 2016, Taiga Nomi
3 All rights reserved.
4
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions are met:
7 * Redistributions of source code must retain the above copyright
8 notice, this list of conditions and the following disclaimer.
9 * Redistributions in binary form must reproduce the above copyright
10 notice, this list of conditions and the following disclaimer in the
11 documentation and/or other materials provided with the distribution.
12 * Neither the name of the <organization> nor the
13 names of its contributors may be used to endorse or promote products
14 derived from this software without specific prior written permission.
15
16 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY
17 EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
18 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
19 DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY
20 DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
21 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
22 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
23 ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
24 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
25 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26*/
27#pragma once
28#include "tiny_dnn/util/util.h"
29#include "tiny_dnn/layers/layer.h"
30
31namespace tiny_dnn {
32
44class concat_layer : public layer {
45public:
49 concat_layer(const std::vector<shape3d>& in_shapes)
50 : layer(std::vector<vector_type>(in_shapes.size(), vector_type::data), {vector_type::data}),
51 in_shapes_(in_shapes) {
52 set_outshape();
53 }
54
59 concat_layer(serial_size_t num_args, serial_size_t ndim)
60 : layer(std::vector<vector_type>(num_args, vector_type::data), { vector_type::data }),
61 in_shapes_(std::vector<shape3d>(num_args, shape3d(ndim,1,1))) {
62 set_outshape();
63 }
64
65 void set_outshape() {
66 out_shape_ = in_shapes_.front();
67 for (size_t i = 1; i < in_shapes_.size(); i++) {
68 if (in_shapes_[i].area() != out_shape_.area())
69 throw nn_error("each input shapes to concat must have same WxH size");
70 out_shape_.depth_ += in_shapes_[i].depth_;
71 }
72 }
73
74 std::string layer_type() const override {
75 return "concat";
76 }
77
78 std::vector<shape3d> in_shape() const override {
79 return in_shapes_;
80 }
81
82 std::vector<shape3d> out_shape() const override {
83 return {out_shape_};
84 }
85
86 void forward_propagation(const std::vector<tensor_t*>& in_data,
87 std::vector<tensor_t*>& out_data) override {
88 serial_size_t num_samples = static_cast<serial_size_t>((*out_data[0]).size());
89
90 for (serial_size_t s = 0; s < num_samples; s++) {
91 float_t* outs = &(*out_data[0])[s][0];
92
93 for (serial_size_t i = 0; i < in_shapes_.size(); i++) {
94 const float_t* ins = &(*in_data[i])[s][0];
95 serial_size_t dim = in_shapes_[i].size();
96 outs = std::copy(ins, ins + dim, outs);
97 }
98 }
99 }
100
101 void back_propagation(const std::vector<tensor_t*>& in_data,
102 const std::vector<tensor_t*>& out_data,
103 std::vector<tensor_t*>& out_grad,
104 std::vector<tensor_t*>& in_grad) override {
105 CNN_UNREFERENCED_PARAMETER(in_data);
106 CNN_UNREFERENCED_PARAMETER(out_data);
107
108 size_t num_samples = (*out_grad[0]).size();
109
110 for (size_t s = 0; s < num_samples; s++) {
111 const float_t* outs = &(*out_grad[0])[s][0];
112
113 for (serial_size_t i = 0; i < in_shapes_.size(); i++) {
114 serial_size_t dim = in_shapes_[i].size();
115 float_t* ins = &(*in_grad[i])[s][0];
116 std::copy(outs, outs + dim, ins);
117 outs += dim;
118 }
119 }
120 }
121
122 template <class Archive>
123 static void load_and_construct(Archive & ar, cereal::construct<concat_layer> & construct) {
124 std::vector<shape3d> in_shapes;
125
126 ar(cereal::make_nvp("in_size", in_shapes));
127 construct(in_shapes);
128 }
129
130 template <class Archive>
131 void serialize(Archive & ar) {
132 layer::serialize_prolog(ar);
133 ar(in_shapes_);
134 }
135
136private:
137 std::vector<shape3d> in_shapes_;
138 shape3d out_shape_;
139};
140
141} // namespace tiny_dnn
concat N layers along depth
Definition concat_layer.h:44
void back_propagation(const std::vector< tensor_t * > &in_data, const std::vector< tensor_t * > &out_data, std::vector< tensor_t * > &out_grad, std::vector< tensor_t * > &in_grad) override
return delta of previous layer (delta=\frac{dE}{da}, a=wx in fully-connected layer)
Definition concat_layer.h:101
std::vector< shape3d > in_shape() const override
array of input shapes (width x height x depth)
Definition concat_layer.h:78
std::string layer_type() const override
name of layer, should be unique for each concrete class
Definition concat_layer.h:74
std::vector< shape3d > out_shape() const override
array of output shapes (width x height x depth)
Definition concat_layer.h:82
void forward_propagation(const std::vector< tensor_t * > &in_data, std::vector< tensor_t * > &out_data) override
Definition concat_layer.h:86
concat_layer(const std::vector< shape3d > &in_shapes)
Definition concat_layer.h:49
concat_layer(serial_size_t num_args, serial_size_t ndim)
Definition concat_layer.h:59
Simple image utility class.
Definition image.h:94
base class of all kind of NN layers
Definition layer.h:62