CHRONONETModel.py

from keras.models import Model
from keras.layers import Input, Dense, concatenate, GRU, Conv1D


def block(input):
    conv1 = Conv1D(32, 2, strides=2, activation='relu', padding="same")(input)
    conv2 = Conv1D(32, 4, strides=2, activation='relu', padding="causal")(input)
    conv3 = Conv1D(32, 8, strides=2, activation='relu', padding="causal")(input)
    x = concatenate([conv1, conv2, conv3], axis=2)
    return x


def gru_model(inputShape):
    input = Input(shape=inputShape)
    block1 = block(input)
    block2 = block(block1)
    block3 = block(block2)

    gru_out1 = GRU(32, activation='tanh', return_sequences=True)(block3)
    gru_out2 = GRU(32, activation='tanh', return_sequences=True)(gru_out1)
    gru_out = concatenate([gru_out1, gru_out2], axis=2)
    gru_out3 = GRU(32, activation='tanh', return_sequences=True)(gru_out)
    gru_out = concatenate([gru_out1, gru_out2, gru_out3])
    gru_out4 = GRU(32, activation='tanh')(gru_out)

    predictions = Dense(1, activation='sigmoid')(gru_out4)
    model = Model(inputs=input, outputs=predictions)

    return model