Models and pre-trained weights (TorchVision)
torchvision.models
Torchvision에서는 몇 가지 모델 구조들에 대한 pre-trained weights를 제공하고 있다.
여러가지 모델들 중 VGG16을 가져다 쓴다면,
1
2
3
4
5
6
from torchvision.models import vgg16_bn, VGG16_BN_Weights
from copy import deepcopy
import math
pretrained_vgg_16 = vgg16_bn(weights=VGG16_BN_Weights.DEFAULT)
full_vgg = deepcopy(pretrained_vgg_16.features[:])
이 상태에서 full_vgg를 출력해보면 다음과 같다.
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
Sequential(
(0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
(3): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(4): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(5): ReLU(inplace=True)
(6): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(7): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(8): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(9): ReLU(inplace=True)
(10): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(11): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(12): ReLU(inplace=True)
(13): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(14): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(15): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(16): ReLU(inplace=True)
(17): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(18): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(19): ReLU(inplace=True)
(20): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(21): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(22): ReLU(inplace=True)
(23): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(24): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(25): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(26): ReLU(inplace=True)
(27): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(28): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(29): ReLU(inplace=True)
(30): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(31): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(32): ReLU(inplace=True)
(33): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
(34): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(35): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(36): ReLU(inplace=True)
(37): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(38): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(39): ReLU(inplace=True)
(40): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(41): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(42): ReLU(inplace=True)
(43): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)
이제 full_vgg의 각 레이어들은 인덱싱이나 슬라이싱이 가능하다.
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
# full_vgg[10]
Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
# full_vgg[10].weight.shape
torch.Size([128, 128, 3, 3])
# full_vgg[10].weight[0]
Parameter containing:
ttensor([[[-0.0189, -0.0044, -0.0155],
[-0.0071, 0.0066, 0.0012],
[ 0.0211, 0.0375, 0.0270]],
[[-0.0474, -0.0388, -0.0478],
[-0.0404, -0.0436, -0.0458],
[-0.0353, -0.0491, -0.0395]],
[[ 0.0035, 0.0046, 0.0037],
[ 0.0043, 0.0052, 0.0044],
[ 0.0037, 0.0045, 0.0039]],
...,
[[ 0.0092, 0.0127, -0.0115],
[-0.0162, -0.0017, -0.0238],
[-0.0105, -0.0083, -0.0093]],
[[-0.0368, -0.0303, -0.0325],
[-0.0324, -0.0234, -0.0286],
[-0.0261, -0.0190, -0.0227]],
[[-0.0063, -0.0425, 0.0260],
[-0.0130, 0.0007, -0.0175],
[-0.0087, 0.0012, -0.0339]]], grad_fn=<SelectBackward0>)
2024.05.30
This post is licensed under CC BY 4.0 by the author.