forked from moon-hotel/DeepLearningWithMe
-
Notifications
You must be signed in to change notification settings - Fork 0
/
02_CnnOpMultiFilters.py
43 lines (40 loc) · 1.87 KB
/
02_CnnOpMultiFilters.py
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
import torch.nn.functional as F
import torch
filter = torch.tensor([[[[0, 0, 1],
[0, 1, 1],
[1, 0, -1]]],
[[[1, 0, 0],
[0, 0, 1],
[1, 0, 1]]]])
inputs = torch.tensor([1, 2, 0, -1, 0, 1, 1, 0, 0, 1, 2, -1, 2, 1, 0,
-1, 1, 0, 0, 0, 2, 1, -1, 0, 0]).reshape([1, 1, 5, 5])
bias = torch.tensor([1, 2])
print(F.conv2d(inputs, filter, padding=0, bias=bias))
filters = torch.tensor([[[[2, 0, 0],
[1, 0, 1],
[0, 3, 0]],
[[1, 0, 1],
[0, 0, 0],
[1, 1, 0]],
[[0, 0, 1],
[1, 1, 1],
[1, 1, 0]]],
[[[0, 1, 0],
[1, 1, 1],
[0, 1, 0]],
[[0, 1, 0],
[1, 0, 1],
[0, 1, 0]],
[[1, 0, 1],
[0, 1, 0],
[1, 0, 1]]]
]) # [2,3,3,3] [ [filter_nums/output_channels,input_channels,high,width]
print("卷积核的形状:[filter_nums/output_channels,input_channels,high,width]", filters.shape)
inputs = torch.tensor([0, 2, 0, 1, 0, 0, 2, 0, 1, 1, 2, 1, 2, 0, 0, 1, 0, 0, 1, 0, -1, 1, 1, 0, 1,
0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0,
# [batch_size,in_channels,high,width]
1, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0]).reshape([1, 3, 5, 5])
bias = torch.tensor([1, -3])
result = F.conv2d(inputs, filters, padding=0, bias=bias)
print("卷积后的结果:", result)
print("结果的形状:", result.shape)