GRU_Model.py

from sklearn.model_selection import train_test_split
from tensorflow import keras as keras
from keras.layers import GRU
from keras.layers import BatchNormalization
from keras.layers import Dense
from keras.layers import Dropout
from keras.layers import Activation
import os
import matplotlib.pyplot as plt

class GRU_Model:
    
    def __init__(self,inputs,targets):    
        os.environ['CUDA_VISIBLE_DEVICES'] = '-1'
        self.INPUT = inputs
        self.OUTPUT = targets
        self.OUTPUT_UNITS = 88
        self.LOSS_FUNCTION = 'sparse_categorical_crossentropy'
        self.LEARNING_RATE = 0.001
        self.BATCH_SIZE = 128
     
    def build_model(self):
        model = keras.models.Sequential()
        model.add(GRU(
            512,
            input_shape = (self.INPUT.shape[1],self.INPUT.shape[2]),
            recurrent_dropout=0,
            return_sequences=True
        ))
        model.add(GRU(512,return_sequences=True,recurrent_dropout=0))
        model.add(GRU(512))
        model.add(BatchNormalization())
        model.add(Dropout(0.3))
        model.add(Dense(256))
        model.add(Activation('relu'))
        model.add(BatchNormalization())
        model.add(Dropout(0.3))
        model.add(Dense(88))
        model.add(Activation('softmax'))

        model.compile(loss=self.LOSS_FUNCTION,
                      optimizer=keras.optimizers.Adam(learning_rate=self.LEARNING_RATE),
                      metrics=['accuracy'])
        
        model.summary()
        return model  


    def train(self):
        model = self.build_model()
        history = model.fit(self.INPUT,self.OUTPUT,epochs=15,batch_size=self.BATCH_SIZE)
        plt.plot(history.history['accuracy'])
        plt.plot(history.history['loss'])
        plt.title('Accuracy vs Loss')
        plt.ylabel('Accuracy')
        plt.xlabel('Epoch')
        plt.legend(['Accuracy','Loss'],loc='upper left')
        plt.show()
        model.save('model_gru.h5')