def __init__(self,
w_initializer,
b_initializer,
v,
lmbda=0,
mask=None,
exclude=None):
super(EqlLayer, self).__init__()
if exclude is None:
exclude = []
self.regularizer = regularizers.L1(l1=lmbda)
self.w_initializer = initializers.get(w_initializer)
self.b_initializer = initializers.get(b_initializer)
self.mask = mask
self.v = v
self.activations = [identity, sin, cos, sigmoid, mult]
self.exclusion = 0
if 'id' in exclude:
self.exclusion += 1
self.activations.remove(identity)
if 'sin' in exclude:
self.exclusion += 1
self.activations.remove(sin)
if 'cos' in exclude:
self.exclusion += 1
self.activations.remove(cos)
if 'sig' in exclude:
self.exclusion += 1
self.activations.remove(sigmoid)
if 'mult' in exclude:
self.exclusion += 2
self.activations.remove(mult)
def model_definition(input_shape):
model = models.Sequential()
model.add(layers.Dense(
units=10,
input_shape=(input_shape,),
use_bias=True,
#activation=activations.relu,
activity_regularizer=regularizers.L2(0.0)))
# model.add(layers.Dense(
# units=3,
# activity_regularizer=regularizers.L2(0.000),
# #activation=activations.relu,
# use_bias=True))
model.add(layers.Dense(
units=1,
activity_regularizer=regularizers.L1(0.0),
#activation=activations.sigmoid,
use_bias=True))
model.compile(
optimizer=optimizers.SGD(0.001),
loss=losses.MSE,
metrics=['acc'])
return model
def initialize_model():
"""NN Model Architecture"""
### Model
model = models.Sequential()
### Embedding Padded
model.add(layers.Embedding(input_dim=5000, output_dim=100, mask_zero=True))
### First convolution & max-pooling
model.add(layers.LSTM(units=100, activation='tanh', return_sequences=True))
model.add(layers.LSTM(units=100, activation='tanh', return_sequences=True))
model.add(layers.LSTM(units=50, activation='tanh'))
model.add(layers.Dropout(0.2)) #change params
model.add(
layers.Dense(
40, activation='relu',
kernel_regularizer=regularizers.L1(0.01))) #Use regulazers
model.add(layers.Dropout(0.2)) #change params
model.add(
layers.Dense(
20, activation='relu',
kernel_regularizer=regularizers.L1(0.01))) #Use regulazers
model.add(layers.Dropout(0.2)) #change params
model.add(
layers.Dense(
10, activation='relu',
kernel_regularizer=regularizers.L1(0.01))) #Use regulazers
model.add(layers.Dropout(0.2)) #change params
### Last layer (let's say a classification with 10 output)
model.add(layers.Dense(1, activation='linear'))
### Model compilation
model.compile(loss='mse', optimizer='rmsprop', metrics=['mae'])
return model
def __init__(self, w_initializer, b_initializer, lmbda=0, mask=None):
super(DenseLayer, self).__init__()
self.regularizer = regularizers.L1(l1=lmbda)
self.w_initializer = initializers.get(w_initializer)
self.b_initializer = initializers.get(b_initializer)
self.mask = mask
t_test = X_test.values
t_y_train = y_train.values
t_y_test = y_test.values
#normalizing the data
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
t_train = scaler.fit_transform(t_train)
t_test = scaler.transform(t_test)
#creating neural network model
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from tensorflow.keras import regularizers
model = Sequential()
model.add(
Dense(74, activation='tanh', kernel_regularizer=regularizers.L1(0.001)))
model.add(
Dense(55, activation='relu', kernel_regularizer=regularizers.L1(0.001)))
model.add(
Dense(40, activation='relu', kernel_regularizer=regularizers.L1(0.001)))
model.add(
Dense(30, activation='relu', kernel_regularizer=regularizers.L1(0.001)))
model.add(Dense(1, activation='relu'))
model.compile(optimizer='adam', loss='mse')
model.fit(x=t_train,
y=t_y_train,
validation_data=(t_test, t_y_test),
batch_size=128,
epochs=340)
loss = pd.DataFrame(model.history.history)
loss.plot()
predictors, label = input_sequences[:, :-1], input_sequences[:, -1]
label = ku.to_categorical(label, num_classes=total_words)
print('Data preparation done')
model = Sequential()
model.add(Embedding(total_words, 100, input_length=max_sequence_len - 1))
# model.add(Conv1D(100, 5, activation='relu')),
# model.add(GlobalAveragePooling1D()),
model.add(Bidirectional(LSTM(150, return_sequences=True)))
model.add(Dropout(0.2))
model.add(Bidirectional(GRU(100, return_sequences=True)))
model.add(LSTM(100))
model.add(
Dense(total_words / 2,
activation='relu',
kernel_regularizer=regularizers.L1(
0.01))) ## A Dense Layer including regularizers
model.add(Dense(total_words, activation='softmax'))
# Pick an optimizer
model.compile(loss='categorical_crossentropy',
optimizer='adam',
metrics=['accuracy'])
print(model.summary())
history = model.fit(predictors, label, epochs=10, verbose=2)
import matplotlib.pyplot as plt
def plot_graphs(history, string):
plt.plot(history.history[string])
plt.xlabel("Epochs")
plt.ylabel(string)
plt.show()