/
model.py
59 lines (43 loc) · 1.89 KB
/
model.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
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Conv2D, BatchNormalization, Activation, ZeroPadding2D, DepthwiseConv2D, Dense, Flatten, LeakyReLU, MaxPooling2D
from tensorflow.keras import regularizers
from hyperparams import Hyperparams
H = Hyperparams()
class Model(object):
def __init__(self):
return None
def get_model(self):
model = Sequential()
model.add(Conv2D(32, kernel_size=3,
input_shape=(192, 256, 3), data_format='channels_last'))
model.add(LeakyReLU(0.2))
model.add(Conv2D(64, kernel_size=5,
data_format='channels_last'))
model.add(LeakyReLU(0.2))
model.add(MaxPooling2D(pool_size=(1, 1), strides=(1, 1),
data_format='channels_last'))
model.add(Conv2D(64, kernel_size=7, strides=(2, 2),
data_format='channels_last'))
model.add(LeakyReLU(0.2))
model.add(MaxPooling2D(pool_size=(2, 2), data_format='channels_last'))
model.add(Conv2D(128, kernel_size=9, strides=(2, 2),
data_format='channels_last'))
model.add(LeakyReLU(0.2))
model.add(MaxPooling2D(pool_size=(2, 2), data_format='channels_last'))
model.add(Conv2D(256, kernel_size=7, strides=(3, 3),
data_format='channels_last',kernel_regularizer=regularizers.l2(0.001)))
model.add(LeakyReLU(0.2))
model.add(Flatten())
model.add(Dense(512, kernel_regularizer=regularizers.l2(0.001)))
model.add(LeakyReLU(0.2))
model.add(Dense(256, kernel_regularizer=regularizers.l2(0.001)))
model.add(LeakyReLU(0.1))
model.add(Dense(128, kernel_regularizer=regularizers.l2(0.001), activation='relu'))
#model.add(LeakyReLU(0.1))
model.add(Dense(64, kernel_regularizer=regularizers.l2(0.001), activation='relu'))
#model.add(LeakyReLU(0.2))
model.add(Dense(16, kernel_regularizer=regularizers.l2(0.001), activation='relu'))
#model.add(LeakyReLU(0.2))
model.add(Dense(4))
return model