def crossValidation(trainingData,
atr,
days,
DAYS_AHEAD,
intervals=10,
EPOCHS=300,
LEARNING_RATE=0.2):
reshaped_x = np.reshape(trainingData, (len(trainingData), 1, atr * days))
reshaped_y = np.reshape(trainingData[:, 2], (len(trainingData), 1, 1))
x = makeBatch(reshaped_x, intervals)
y = makeBatch(reshaped_y, intervals)
errors = []
for i in range(len(x)):
x_train = getSelection(x, i)[:-DAYS_AHEAD]
y_train = getSelection(y, i)[DAYS_AHEAD:]
x_test = x[i][:-DAYS_AHEAD]
y_test = y[i][DAYS_AHEAD:]
inputSize = 2 * days
outputSize = int(inputSize / 2)
net = Network()
net.add(FCLayer(inputSize, outputSize * 3))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(outputSize * 3, outputSize))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(outputSize, 1))
net.add(ActivationLayer(tanh, tanh_prime))
# train
net.use(mse, mse_prime)
net.fit(x_train, y_train, epochs=EPOCHS, learning_rate=LEARNING_RATE)
# test
out, err = net.predict(x_test, y_test)
errors.append(err)
print(err)
print(np.average(errors))
import numpy as np from network import Network from fc_layer import FCLayer from activation_layer import ActivationLayer #from activations import tanh, tanh_prime from losses import mse, mse_prime from activations import sigmoid, sigmoid_prime # training data x_train = np.array([[[0, 0]], [[0, 1]], [[1, 0]], [[1, 1]]]) y_train = np.array([[[0]], [[1]], [[1]], [[0]]]) # network net = Network() net.add(FCLayer(2, 3)) net.add(ActivationLayer(sigmoid, sigmoid_prime)) net.add(FCLayer(3, 1)) net.add(ActivationLayer(sigmoid, sigmoid_prime)) # train net.use(mse, mse_prime) cost_, myerr = net.fit(x_train, y_train, epochs=10000, learning_rate=0.2) # test out = net.predict(x_train) print(out) import matplotlib.pyplot as plt plt.plot(cost_)
x_train = x_train.reshape(x_train.shape[0], 1, 28 * 28)
x_train = x_train.astype('float32')
x_train /= 255
# encode output which is a number in range [0,9] into a vector of size 10
# e.g. number 3 will become [0, 0, 0, 1, 0, 0, 0, 0, 0, 0]
y_train = np_utils.to_categorical(y_train)
# same for test data : 10000 samples
x_test = x_test.reshape(x_test.shape[0], 1, 28 * 28)
x_test = x_test.astype('float32')
x_test /= 255
y_test = np_utils.to_categorical(y_test)
# Network
net = Network()
net.add(FCLayer(28 * 28,
100)) # input_shape=(1, 28*28) ; output_shape=(1, 100)
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(100, 50)) # input_shape=(1, 100) ; output_shape=(1, 50)
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(50, 10)) # input_shape=(1, 50) ; output_shape=(1, 10)
net.add(ActivationLayer(tanh, tanh_prime))
# train on 1000 samples
# as we didn't implemented mini-batch GD, training will be pretty slow if we update at each iteration on 60000 samples...
net.use(mse, mse_prime)
net.fit(x_train[0:1000], y_train[0:1000], epochs=50, learning_rate=0.01)
# test on 3 samples
out = net.predict(x_test[0:3])
print("\n")
print("predicted values : ")
def visualise_image(data, num_image):
image = np.asarray(data[num_image].squeeze())
plt.imshow(image)
plt.show()
x_train = import_image('train-images-idx3-ubyte.gz', 6000)
x_train /= 255
y_train = import_labels('train-labels-idx1-ubyte.gz', 6000)
x_test = import_image('t10k-images-idx3-ubyte.gz', 1000)
x_test /= 255
y_test = import_labels('t10k-labels-idx1-ubyte.gz', 1000)
y_train = to_categorical(y_train) #copied from tensorflow.keras.np_utils
y_test = to_categorical(y_test)
#x_train = np.asarray(x_train)
#training
net = Network()
net.add(FCLayer(image_size**2,
100)) #input (1, input_shape**2) , output (1,100)
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(100, 50)) #input (1,100 ) , output (1, 50)
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(50, 10)) #input (1,50) , output(1,10)
net.add(ActivationLayer(tanh, tanh_prime))
net.use(mse, mse_prime)
net.fit(x_train[0:1000], y_train[0:1000], epochs=35, learning_rate=0.1)
t2.append(t[k])
return quicksort(t1) + [pivot] + quicksort(t2)
## init
population = 10 #nombre total de cas tester en meme temps
nombre = 1000 #nombre de géneration
q = 0.3 #facteur de mutation (entre 0 et 1)
x_train = np.array([[[0, 0]], [[0, 1]], [[1, 0]],
[[1, 1]]]) #ce qui est en entrée
##init des reseaux de neurones aléatoirement
liste = []
for k in range(0, population): #on crée "population" réseaux de neurones
liste.append([0, Network()])
liste[-1][1].add(FCLayer(2, 3)) # construction des couches
liste[-1][1].add(ActivationLayer(tanh)) #choix fonction d'activation
liste[-1][1].add(FCLayer(3, 3))
liste[-1][1].add(ActivationLayer(tanh))
liste[-1][1].add(FCLayer(3, 1))
liste[-1][1].add(ActivationLayer(tanh))
n = 0
while (n < nombre): #tant que l'on n'est pas à la "nombre"ème génerations
n += 1
for k in range(0, population): #pour tout les réseaux de neurones
y_trouve = []
y_voulu = []
for x in x_train: #on met les 4 entrées
y = liste[k][1].predict(x) #on teste les reseaux de neurones
y_attendu = x[0][0] ^ x[0][1] #la valeur que l'on doit avoir
y_voulu.append(y_attendu) #on met les resultats dans un tableau
x = df[ls[1:-1]].to_numpy()[:, np.newaxis]
y = df[ls[-1]].to_numpy().reshape(-1, 1)
#Split into training and testing data
x_train, x_test, y_train, y_test = train_test_split(x,
y,
test_size=0.33,
random_state=42)
x_train, x_val, y_train, y_val = train_test_split(
x_train, y_train, test_size=0.15,
random_state=42) #Validation data of 10% (15% of 67%)
np.random.seed(42)
#Initialise Neural Network
net = Network()
n_hidden_nodes = 500
#Input layer
net.add(FCLayer(x.shape[-1], n_hidden_nodes))
net.add(ActivationLayer(tanh, tanh_prime))
#hidden layers
net.add(FCLayer(n_hidden_nodes, n_hidden_nodes))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(n_hidden_nodes, n_hidden_nodes))
net.add(ActivationLayer(tanh, tanh_prime))
#Output layer
net.add(FCLayer(n_hidden_nodes, 1))
#Took out activation function here
net.use(mse, mse_prime)
net.fit(x_train, y_train, epochs=100, learning_rate=0.001)
y_pred = np.array(net.predict(x_test)).reshape(-1, 1)
np.savetxt('data_test500Trial.csv', y_pred, delimiter=',')
def test_update_parameters(self):
# test example from coursera deeplearning.ai
np.random.seed(1)
W1 = cp.array(np.random.randn(2, 3))
b1 = cp.array(np.random.randn(2, 1))
W2 = cp.array(np.random.randn(3, 3))
b2 = cp.array(np.random.randn(3, 1))
dW1 = cp.array(np.random.randn(2, 3))
db1 = cp.array(np.random.randn(2, 1))
dW2 = cp.array(np.random.randn(3, 3))
db2 = cp.array(np.random.randn(3, 1))
l1 = FCLayer(3, 2)
l1.W = W1
l1.b = b1
l1.dW = dW1
l1.db = db1
l2 = FCLayer(3, 3)
l2.W = W2
l2.b = b2
l2.dW = dW2
l2.db = db2
l1.update_parameters(0.01, 2)
l2.update_parameters(0.01, 2)
self.assertAlmostEqual(float(l1.W[0, 0]), 1.63178673)
self.assertAlmostEqual(float(l2.sb[0, 0]), 5.49507194e-05)
self.assertAlmostEqual(float(l1.b[1, 0]), -0.75376553)
def test_init_parameters(self):
np.random.seed(3)
layer = FCLayer(2, 4)
def test_backward_propagation(self):
# test example from coursera deeplearning.ai
np.random.seed(3)
AL = cp.array(np.random.randn(1, 2))
Y = cp.array([[1, 0]])
dA = -(Y / AL - (1 - Y) / (1 - AL))
A1 = cp.array(np.random.randn(4, 2))
W1 = cp.array(np.random.randn(3, 4))
b1 = cp.array(np.random.randn(3, 1))
Z1 = cp.array(np.random.randn(3, 2))
A2 = cp.array(np.random.randn(3, 2))
W2 = cp.array(np.random.randn(1, 3))
b2 = cp.array(np.random.randn(1, 1))
Z2 = cp.array(np.random.randn(1, 2))
l1 = FCLayer(4, 3, activation='relu')
l1.Z = Z1
l1.W = W1
l1.b = b1
l1.X = A1
l2 = FCLayer(3, 2, activation='sigmoid')
l2.Z = Z2
l2.W = W2
l2.b = b2
l2.X = A2
dA = l2.backward(dA)
dA = l1.backward(dA)
self.assertAlmostEqual(0.41010002, float(l1.dW[0, 0]), places=7)
self.assertAlmostEqual(0.01005865, float(l1.dW[2, 1]), places=7)
self.assertAlmostEqual(-0.02835349, float(l1.db[2, 0]), places=7)
def test_forward_propagation(self):
# test example from coursera deeplearning.ai
np.random.seed(6)
X = cp.array(np.random.randn(5, 4))
W1 = cp.array(np.random.randn(4, 5))
b1 = cp.array(np.random.randn(4, 1))
W2 = cp.array(np.random.randn(3, 4))
b2 = cp.array(np.random.randn(3, 1))
W3 = cp.array(np.random.randn(1, 3))
b3 = cp.array(np.random.randn(1, 1))
lay1 = FCLayer(5, 4, activation='relu')
lay1.W = W1
lay1.b = b1
lay2 = FCLayer(4, 3, activation='relu')
lay2.W = W2
lay2.b = b2
lay3 = FCLayer(3, 1, activation='sigmoid')
lay3.W = W3
lay3.b = b3
A = lay1.forward(X)
A = lay2.forward(A)
A = lay3.forward(A)
self.assertAlmostEqual(0.03921668, float(A[0, 0]), places=7)
self.assertAlmostEqual(0.19734387, float(A[0, 2]), places=7)
(len(trainingData), 1, ATR * DAYS))[0:SIZE - DAYS_AHEAD]
y_train = np.reshape(trainingData[:, 2],
(len(trainingData), 1, 1))[DAYS_AHEAD:SIZE]
time_train = dataArray[DAYS_AHEAD:SIZE]
x_test = np.reshape(trainingData,
(len(trainingData), 1, ATR * DAYS))[SIZE:-DAYS_AHEAD]
y_test = np.reshape(trainingData[:, 2],
(len(trainingData), 1, 1))[SIZE + DAYS_AHEAD:]
time_test = dataArray[SIZE + DAYS + DAYS_AHEAD:]
inputSize = len(x_train)
# network
inputSize = ATR * DAYS
outputSize = int(inputSize / 2)
net = Network()
net.add(FCLayer(inputSize, outputSize * 3))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(outputSize * 3, outputSize))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(outputSize, 1))
net.add(ActivationLayer(tanh, tanh_prime))
# train
net.use(mse, mse_prime)
net.fit(x_train, y_train, epochs=EPOCHS, learning_rate=LEARNING_RATE)
# test
out, err = net.predict(x_test, y_test)
print(err)
outTrain, err = net.predict(x_train, y_train)
from network import Network
from fc_layer import FCLayer
from softmax_layer import SoftmaxLayer
from activation_layer import ActivationLayer
from activations import tanh, tanh_prime, softmax, softmax_prime
from losses import mse, mse_prime, cross_entropy, cross_entropy_prime
# TRAINING DATA
# input x_train shape is (4,1,2):
x_train = np.array([[[0, 0]], [[0, 1]], [[1, 0]], [[1, 1]]])
#y_train =np.array([ [[0]], [[1]], [[1]], [[0]] ])
y_train = np.array([[1, 0], [0, 1], [0, 1], [1, 0]]) # 1 hot vector
# NETWORK
net = Network()
net.add(FCLayer(2, 3)) # input co 2 dimension [0,0], 3 neron
# FCLayer sử dụng hàm activate là tanh, đạo hàm của tanh là tanh_prime:
net.add(ActivationLayer(tanh, tanh_prime))
#net.add(FCLayer(3, 1)) # input co 3 dimension, 1 neron
net.add(FCLayer(3, 2)) # input co 3 dimension, 2 neron
#net.add(ActivationLayer(tanh, tanh_prime))
net.add(ActivationLayer(softmax, softmax_prime))
"""
Thêm lớp ở trên thì bị lỗi sau:
Traceback (most recent call last):
File "example_xor.py", line 36, in <module>
net.fit(x_train, y_train, epochs=1000, learning_rate=0.1)
File "E:\MyProg\Python\medium_nn\network.py", line 58, in fit
error = layer.backward_propagation(error, learning_rate)
File "E:\MyProg\Python\medium_nn\fc_layer.py", line 29, in backward_propagation
weights_error = np.dot(self.input.T, output_error)
x_train /= 255
#ecnoding output
y_train = np_utils.to_categorical(y_train)
#same for test data
x_test = x_test.reshape(x_test.shape[0], 1, 28 * 28)
x_test = x_test.astype('float32')
x_test /= 255
#ecnoding output
y_test = np_utils.to_categorical(y_test)
#Network
net = Network()
net.add(FCLayer(28 * 28, 50))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(50, 10))
net.add(ActivationLayer(tanh, tanh_prime))
net.use(mse, mse_prime)
net.fit(x_train[:1000], y_train[:1000], epochs=50, learning_rate=0.1)
#test on 3 samples
out = net.predict(x_test[:1])
print('\n')
print('true values: ')
print(y_test[0:1])
print('\n')
print('predicted values: ')
y_train = np_utils.to_categorical(y_train)
# same for test data : 10000 samples
x_test = x_test.reshape(x_test.shape[0], 28, 28, 1)
x_test = x_test.astype('float32')
x_test /= 255
y_test = np_utils.to_categorical(y_test)
# Network
net = Network()
net.add(ConvLayer((28, 28, 1), (3, 3),
1)) # input_shape=(28, 28, 1) ; output_shape=(26, 26, 1)
net.add(ActivationLayer(tanh, tanh_prime))
net.add(
FlattenLayer()) # input_shape=(26, 26, 1) ; output_shape=(1, 26*26*1)
net.add(FCLayer(26 * 26 * 1,
100)) # input_shape=(1, 26*26*1) ; output_shape=(1, 100)
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(100, 10)) # input_shape=(1, 100) ; output_shape=(1, 10)
net.add(ActivationLayer(tanh, tanh_prime))
# train on 1000 samples
# as we didn't implemented mini-batch GD, training will be pretty slow if we update at each iteration on 60000 samples...
net.use(mse, mse_prime)
net.fit(x_train[0:1000], y_train[0:1000], epochs=100, learning_rate=0.1)
# test on 3 samples
out = net.predict(x_test[0:3])
print("\n")
print("predicted values : ")
print(out, end="\n")
print("true values : ")
[[0, 0]],
[[0, 1]],
[[1, 0]],
[[1, 1]],
[[2, 0]],
[[2, 1]],
[[3, 0]],
[[3, 1]]
])
# New state
y_train = np.array([[[1]], [[3]], [[2]], [[0]], [[3]], [[1]], [[0]], [[2]]])
# Network
net = Network()
net.add(FCLayer(2, 4))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(4, 1))
# Train
net.use(mse, mse_prime)
net.fit(x_train, y_train, epochs=4000, learning_rate=0.02)
# Test
out = net.predict(x_train)
x = np.round(out)
# Take user input and perform state transitions
user_input = int(input("Enter the input (0 or 1 or -1):"))
current_state = 0
print("Current state is", current_state)
x_test = x_norm.to_numpy()
y_true = data_test.drop(columns=[
'Index', '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'
]).to_numpy()
y_true = np.where(y_true == 'M', 1, 0)
y = data_train.drop(columns=[
'Index', '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'
]).to_numpy()
y_train = np.where(y == 'M', 1, 0)
# network
net = Network()
net.add(FCLayer(30, 15))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(15, 10))
net.add(ActivationLayer(tanh, tanh_prime))
net.add(FCLayer(10, 1))
net.add(ActivationLayer(tanh, tanh_prime))
#train
net.use(mse, mse_prime)
net.fit(x_train, y_train, epochs=100, learning_rate=0.1)
#test
out = net.predict(x_test)
print(out)
y_pred = [1 if x > 0.5 else 0 for x in out]
print("accuracy : ", accuracy_score_(y_pred, y_true))
else:
print("Usage : python Network.py train.csv test.csv")