Alexnet.cnna 1014 Bytes
Newer Older
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
architecture Alexnet{
    def input image
    def output predictions

    image ->
    convolution(filters=96, kernel=11, stride=4)
    activation.relu()
    pooling.max(kernel=3, stride=2)
    lrn(nsize=5, alpha=0.0001, beta=0.75)
    
    group 1..2{
        split()
        convolution(filters=128, kernel=5)
        activation.relu()
        pooling.max(kernel=3, stride=2)
        lrn(nsize=5, alpha=0.0001, beta=0.75)
    } merge concatenate()
    
    convolution(filters=384, kernel=3)
    activation.relu()
    
    group 1..2{
        split()
        convolution(filters=192, kernel=3)
        activation.relu()
        
        convolution(filters=128, kernel=3)
        activation.relu()
        pooling.max(kernel=3, stride=2)
    } merge concatenate()
    
    fullyConnected(units=4096)
    activation.relu()
    dropout(p=0.5)
    
    fullyConnected(units=4096)
    activation.relu()
    dropout(p=0.5)
    
    output{
        fullyConnected()
        activation.softmax()
    } -> predictions
}