def main3():
net = Network(cost="categorical_cross_entropy")
net.append(Dense(2, 5, activation="sigmoid"))
net.append(Dense(5, 3, activation="softmax"))
train_X = np.asarray([[0.2, -0.3], [0.6, -0.2], [0.8, 0.9], [0.1, 0.1]])
train_Y = np.asarray([[0.0, 1.0, 0.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0], [1.0, 0.0, 0.0]])
sgd = SGD(alpha=0.05, n_epochs=20, mini_batch_size=4, verbosity=1)
net.fit(train_X, train_Y, sgd)
def main2():
net = Network(cost="quadratic")
net.append(Dense(2, 5))
net.append(Dense(5, 3))
train_X = np.asarray([[0.2, -0.3], [0.6, -0.2], [0.8, 0.9], [0.1, 0.1]])
train_Y = np.asarray([[0.0, 1.0, 0.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0], [1.0, 0.0, 0.0]])
sgd = SGD(alpha=0.1, n_epochs=10, mini_batch_size=3)
net.fit(train_X, train_Y, sgd)
def main3():
net = Network(cost="categorical_cross_entropy")
net.append(Dense(2, 5, activation="sigmoid"))
net.append(Dense(5, 3, activation="softmax"))
train_X = np.asarray([[0.2, -0.3], [0.6, -0.2], [0.8, 0.9], [0.1, 0.1]])
train_Y = np.asarray([[0.0, 1.0, 0.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[1.0, 0.0, 0.0]])
sgd = SGD(alpha=0.05, n_epochs=20, mini_batch_size=4, verbosity=1)
net.fit(train_X, train_Y, sgd)
def main2():
net = Network(cost="quadratic")
net.append(Dense(2, 5))
net.append(Dense(5, 3))
train_X = np.asarray([[0.2, -0.3], [0.6, -0.2], [0.8, 0.9], [0.1, 0.1]])
train_Y = np.asarray([[0.0, 1.0, 0.0], [0.0, 0.0, 1.0], [0.0, 1.0, 0.0],
[1.0, 0.0, 0.0]])
sgd = SGD(alpha=0.1, n_epochs=10, mini_batch_size=3)
net.fit(train_X, train_Y, sgd)
