DMVST-Net/model.py at master · hbezlls/DMVST-Net · GitHub

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
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
import tensorflow as tf
from utils import *
from keras import backend as K
import numpy as np
import pandas as pd
import argparse
from keras.models import Sequential, Model
from keras import activations
from keras.engine.topology import Layer, InputSpec
from keras.utils import conv_utils
from keras.layers import LSTM, InputLayer, Dense, Input, Flatten, concatenate, Reshape
from keras.callbacks import EarlyStopping
from sklearn.preprocessing import MinMaxScaler
from sklearn.metrics import mean_squared_error
from keras.optimizers import Adam
from keras import metrics
from keras.layers.normalization import BatchNormalization
from random import randint
import pickle

batch_size = 64
mean_label = 0.0
label_max = 0
label_min = 0
max_epoch = 200
num_feature = 100
seq_len = 8
hidden_dim = 512
threshold = 10.0
maxtruey = 0
mintruey = 0
eps = 1e-5
loss_lambda = 10.0
feature_len = 0
local_image_size = 9
cnn_hidden_dim_first = 32
len_valid_id = 0
toponet_len = 32

sess = tf.Session()
K.set_session(sess)


class Local_Seq_Conv(Layer):

    def __init__(self, output_dim, seq_len, feature_size, kernel_size, activation=None,
                 kernel_initializer='glorot_uniform',
                 bias_initializer='zeros', padding='same', strides=(1, 1), **kwargs):
        super(Local_Seq_Conv, self).__init__(**kwargs)
        self.output_dim = output_dim
        self.seq_len = seq_len
        self.bias_initializer = bias_initializer
        self.kernel_size = kernel_size
        self.kernel_initializer = kernel_initializer
        self.padding = padding
        self.strides = strides
        self.activation = activations.get(activation)

    def build(self, input_shape):
        batch_size = input_shape[0]
        self.kernel = []
        self.bias = []
        for eachlen in range(self.seq_len):
            self.kernel += [self.add_weight(shape=self.kernel_size,
                                            initializer=self.kernel_initializer,
                                            trainable=True, name='kernel_{0}'.format(eachlen))]

            self.bias += [self.add_weight(shape=(self.kernel_size[-1],),
                                          initializer=self.bias_initializer,
                                          trainable=True, name='bias_{0}'.format(eachlen))]
        self.build = True

    def call(self, inputs):
        output = []
        for eachlen in range(self.seq_len):

            tmp = K.bias_add(K.conv2d(inputs[:, eachlen, :, :, :], self.kernel[eachlen],
                                      strides=self.strides, padding=self.padding), self.bias[eachlen])

            if self.activation is not None:
                output += [self.activation(tmp)]

        output = tf.stack(output, axis=1)
        return output

    def compute_output_shape(self, input_shape):
        return (input_shape[0], input_shape[1], input_shape[2], input_shape[3], self.output_dim)


def build_model(trainX, trainY, testX, testY, trainimage, testimage, traintopo, testtopo, feature_len):
    # X_train, Y_train, X_test, Y_test = Featureset_get()
    image_input = Input(shape=(seq_len, local_image_size,
                               local_image_size, None), name='cnn_input')
    spatial = Local_Seq_Conv(output_dim=cnn_hidden_dim_first, seq_len=seq_len, feature_size=feature_len,
                             kernel_size=(3, 3, 1, cnn_hidden_dim_first), activation='relu',
                             kernel_initializer='glorot_uniform', bias_initializer='zeros', padding='same',
                             strides=(1, 1))(image_input)
    spatial = BatchNormalization()(spatial)
    # spatial = Local_Seq_Pooling(seq_len=seq_len)(spatial)
    spatial = Local_Seq_Conv(output_dim=cnn_hidden_dim_first, seq_len=seq_len, feature_size=feature_len,
                             kernel_size=(3, 3, cnn_hidden_dim_first, cnn_hidden_dim_first), activation='relu',
                             kernel_initializer='glorot_uniform', bias_initializer='zeros', padding='same',
                             strides=(1, 1))(spatial)
    spatial = BatchNormalization()(spatial)
    # spatial = Local_Seq_Pooling(seq_len=seq_len)(spatial)
    spatial = Local_Seq_Conv(output_dim=cnn_hidden_dim_first, seq_len=seq_len, feature_size=feature_len,
                             kernel_size=(3, 3, cnn_hidden_dim_first, cnn_hidden_dim_first), activation='relu',
                             kernel_initializer='glorot_uniform', bias_initializer='zeros', padding='same',
                             strides=(1, 1))(spatial)
    # spatial = Local_Seq_Pooling(seq_len=seq_len)(spatial)
    # spatial = BatchNormalization()(spatial)
    # spatial = Local_Seq_Conv(output_dim=cnn_hidden_dim_first, seq_len=seq_len, feature_size=feature_len, kernel_size=(3, 3, cnn_hidden_dim_first, cnn_hidden_dim_first), activation='relu', kernel_initializer='glorot_uniform', bias_initializer='zeros', padding='same', strides=(1, 1))(spatial)
    # spatial = BatchNormalization()(spatial)
    # spatial = Local_Seq_Pooling(seq_len=seq_len)(spatial)
    # spatial = Local_Seq_Conv(output_dim=cnn_hidden_dim_first, seq_len=seq_len, feature_size=feature_len, kernel_size=(3, 3, cnn_hidden_dim_first, cnn_hidden_dim_first), activation='relu', kernel_initializer='glorot_uniform', bias_initializer='zeros', padding='same', strides=(1, 1))(spatial)
    spatial = Flatten()(spatial)
    spatial = Reshape(target_shape=(seq_len, -1))(spatial)
    spatial_out = Dense(units=64, activation='relu')(spatial)

    lstm_input = Input(shape=(seq_len, feature_len),
                       dtype='float32', name='lstm_input')

    x = concatenate([lstm_input, spatial_out], axis=-1)
    # lstm_out = Dense(units=128, activation=relu)(x)
    lstm_out = LSTM(units=hidden_dim, return_sequences=False, dropout=0)(x)

    topo_input = Input(shape=(toponet_len,), dtype='float32', name='topo_input')
    topo_emb = Dense(units=6, activation='tanh')(topo_input)
    static_dynamic_concate = concatenate([lstm_out, topo_emb], axis=-1)

    res = Dense(units=1, activation='sigmoid')(static_dynamic_concate)
    model = Model(inputs=[image_input, lstm_input, topo_input],
                  outputs=res)
    sgd = Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=1e-6)
    model.compile(loss=mean_absolute_percentage_error_revise, optimizer=sgd,
                  metrics=[metrics.mae])
    earlyStopping = EarlyStopping(
        monitor='val_loss', patience=10, verbose=0, mode='min')
    model.fit([trainimage, trainX, traintopo], trainY, batch_size=batch_size, epochs=max_epoch, validation_split=0.1,
              callbacks=[earlyStopping])
    # model.save('local_conv_lstm_total_embed.h5')
    return model