Inheritance diagram for resnet2d.ResNet2d:
Public Member Functions | |
| None | __init__ (self, in_channels, stage_blocks, input_key) |
| forward (self, sample) | |
Public Attributes | |
| inplanes | |
| input_key | |
| conv1 | |
| bn1 | |
| relu | |
| layer1 | |
| layer2 | |
| layer3 | |
| layer4 | |
Protected Member Functions | |
| nn.Sequential | _make_layer (self, block, int planes, int blocks, int stride=1) |
Detailed Description
이미지 특징을 사용하기 위한 모델 입력을 T x 1 로 가정하고 만듦
Definition at line 42 of file resnet2d.py.
Constructor & Destructor Documentation
◆ __init__()
| None resnet2d.ResNet2d.__init__ | ( | self, | |
| in_channels, | |||
| stage_blocks, | |||
| input_key ) |
args:
in_channels (int) : 입력 채널수
stage_blocks (List) : 스테이지 마다 쌓는 블럭의 수를 설정
input_key (str) : 모듈의 inference에서 사용하는 입력데이터의 키값
Definition at line 47 of file resnet2d.py.
52 ) -> None:
53 """
54 args:
55 in_channels (int) : 입력 채널수
56 stage_blocks (List) : 스테이지 마다 쌓는 블럭의 수를 설정
57 input_key (str) : 모듈의 inference에서 사용하는 입력데이터의 키값
58
59 """
60 super(ResNet2d, self).__init__()
61
62 block = BasicBlock
63
64 self.inplanes = in_channels
65 self.input_key = input_key
66
67 # input block
68 self.conv1 = nn.Conv2d(in_channels, self.inplanes, kernel_size=1, stride=1, padding=0, bias=True)
69 self.bn1 = nn.BatchNorm2d(self.inplanes)
70 self.relu = nn.ReLU(inplace=True)
71
72 # residual blocks
73 self.layer1 = self._make_layer(block, in_channels, stage_blocks[0], stride=1)
74 self.layer2 = self._make_layer(block, in_channels, stage_blocks[1], stride=2)
75 self.layer3 = self._make_layer(block, in_channels, stage_blocks[2], stride=2)
76 self.layer4 = self._make_layer(block, in_channels, stage_blocks[3], stride=1)
77
78
Member Function Documentation
◆ _make_layer()
|
protected |
Definition at line 79 of file resnet2d.py.
79 def _make_layer(self, block, planes: int, blocks: int, stride: int = 1) -> nn.Sequential:
80
81 downsample = None
82
83 # downsampling 필요할경우 downsample layer 생성
84 if stride != 1 or self.inplanes != planes:
85 downsample = nn.Sequential(
86 nn.Conv2d(self.inplanes, planes, kernel_size=1, stride=stride),
87 nn.BatchNorm2d(planes)
88 )
89
90 layers = []
91 layers.append(block(self.inplanes, planes, stride, downsample))
92 self.inplanes = planes
93 for _ in range(1, blocks):
94 layers.append(block(self.inplanes, planes))
95
96 return nn.Sequential(*layers)
97
◆ forward()
| resnet2d.ResNet2d.forward | ( | self, | |
| sample ) |
args:
sample (dict)) : 입력 데이터, self.input_key에 해당하는 키가 있어야함
self.input_key의 아이템은 Tensor 타입 -> shape (B, C, T, 1)
B : 배치 크기
C : 입력 채널
T : 시간
return (Tensor):
특정 해상도의 특징 벡터 -> shape (B, C_o, T_o, 1)
B : 배치 크기
C_o : 채널
T_o : 시간
Definition at line 98 of file resnet2d.py.
98 def forward(self, sample):
99 """
100 args:
101 sample (dict)) : 입력 데이터, self.input_key에 해당하는 키가 있어야함
102 self.input_key의 아이템은 Tensor 타입 -> shape (B, C, T, 1)
103 B : 배치 크기
104 C : 입력 채널
105 T : 시간
106
107 return (Tensor):
108 특정 해상도의 특징 벡터 -> shape (B, C_o, T_o, 1)
109 B : 배치 크기
110 C_o : 채널
111 T_o : 시간
112 """
113 x = sample[self.input_key]
114 x = self.conv1(x)
115 x = self.bn1(x)
116 x = self.relu(x)
117
118 x = self.layer1(x)
119 x = self.layer2(x)
120 x = self.layer3(x)
121 x = self.layer4(x)
122
123 return x
Member Data Documentation
◆ bn1
| resnet2d.ResNet2d.bn1 |
Definition at line 69 of file resnet2d.py.
◆ conv1
| resnet2d.ResNet2d.conv1 |
Definition at line 68 of file resnet2d.py.
◆ inplanes
| resnet2d.ResNet2d.inplanes |
Definition at line 64 of file resnet2d.py.
◆ input_key
| resnet2d.ResNet2d.input_key |
Definition at line 65 of file resnet2d.py.
◆ layer1
| resnet2d.ResNet2d.layer1 |
Definition at line 73 of file resnet2d.py.
◆ layer2
| resnet2d.ResNet2d.layer2 |
Definition at line 74 of file resnet2d.py.
◆ layer3
| resnet2d.ResNet2d.layer3 |
Definition at line 75 of file resnet2d.py.
◆ layer4
| resnet2d.ResNet2d.layer4 |
Definition at line 76 of file resnet2d.py.
◆ relu
| resnet2d.ResNet2d.relu |
Definition at line 70 of file resnet2d.py.
The documentation for this class was generated from the following file:
- smaction/models/backbones/resnet2d.py
