model.py

#!/opt/anaconda3/bin/python

import tensorflow as tf
import numpy

from fourierWeight import fourierLayer, fourierLayerShape, fourierWeight
import tensorflow.keras
from keras.models import Model
from keras.regularizers import l2
from keras.optimizers import Adam

from keras.layers import (
    Input,
    Dense,
    # Convolution2D,
    Conv2D,
    Lambda,
    Activation,
    Flatten,
    Dropout,
    Bidirectional,
    LSTM,
    Reshape,
    TimeDistributed
)

DROP_RATE = 0.5
L2_RATE = 0.0001


def KerasModel(isCompile=True):
    In = Input(shape=(None, 1, 1))
    x = Lambda(fourierLayer, output_shape=fourierLayerShape)(In)
    #x = Convolution2D(32, 7, 7, subsample=(3,3), activation='relu', W_regularizer=l2(L2_RATE))(x)

    x = Conv2D(32, (7, 7), activation='relu', strides=(
        3, 3), kernel_regularizer=l2(L2_RATE))(x)

    #x=Conv2D(32, kernel_size=(7, 7), activation='relu')(x)
    #x = Dropout(DROP_RATE)(x)
    #x = Convolution2D(64, 7, 5, subsample=(3,3), activation='relu', W_regularizer=l2(L2_RATE))(x)
    x = Conv2D(64, (7, 5), activation='relu', strides=(
        3, 3), kernel_regularizer=l2(L2_RATE))(x)
    #x=Conv2D(64, kernel_size=(7, 5), activation='relu')(x)
    #x = Dropout(DROP_RATE)(x)
    #x = Convolution2D(64, 3, 3, subsample=(2,2), activation='relu', W_regularizer=l2(L2_RATE))(x)
    x = Conv2D(64, (2, 2), activation='relu', strides=(
        2, 2), kernel_regularizer=l2(L2_RATE))(x)
    #x=Conv2D(64, kernel_size=(3, 3), activation='relu')(x)
    #x = Dropout(DROP_RATE)(x)
    #x = Convolution2D(32, 3, 3, subsample=(2,2), activation='relu', W_regularizer=l2(L2_RATE))(x)
    x = Conv2D(32, (2, 2), activation='relu', strides=(
        2, 2), kernel_regularizer=l2(L2_RATE))(x)
    # x=Conv2D(32, kernel_size=(3, 3), activation='relu')(x)
    #x = Dropout(DROP_RATE)(x)

    freq, chan = x.get_shape()[2:4]
    x = TimeDistributed(Reshape([int(freq)*int(chan)]))(x)

    x = Bidirectional(LSTM(128, ))(x)

    #x = Dropout(DROP_RATE)(x)
    x = Dense(64, activation='relu')(x)
    #x = Dropout(DROP_RATE)(x)
    x = Dense(24, activation='softmax')(x)

    model = Model(inputs=In, outputs=x)

    if isCompile:
        opt = Adam(1e-4)
        model.compile(
            optimizer=opt,
            loss='categorical_crossentropy',
            metrics=['accuracy'],
        )

    return model


if __name__ == '__main__':
    # model()
    KerasModel()