def test_cost_function(train_data, test_data):
x, y, y_onehot, = train_data
x_test, y_test = test_data
network = NeuralNetwork(layers=[784, 25, 10])
network.fit(x, y_onehot, alpha=0.1, iterations=40)
prediction_test = network.predict(x_test)
accuracy_test = accuracy(prediction_test, y_test)
assert accuracy_test > 90
model.add_layer(Layer(8, 3))
# 构建损失函数和优化器
lr = 0.01
loss = Loss(loss='cross_entropy_with_logits')
optimizer = Optimizers(optimizer='sgd', learning_rate=lr)
model.compile(loss=loss, optimizer=optimizer)
# 训练数据
num_epochs = 1600
batch_size = 64
train_loss = []
test_loss = []
for epoch in range(num_epochs):
for x, y in train_data.get_batch(batch_size):
loss = model.fit(x, y)
train_loss.append(loss)
t_loss, n, right_num = 0., 0, 0
for x, y in test_data.get_batch(batch_size, shuffle=False):
y_pred = model(x)
right_num += np.sum(
np.argmax(y_pred, axis=-1) == np.argmax(y, axis=-1))
t_loss += model.loss(y_pred, y)
n += 1
accuracy = right_num / test_data.num_samples
t_loss /= n
test_loss.append(t_loss)
print('>%d/%d: train loss=%f, test_loss=%f, accuracy=%f' %
(epoch + 1, num_epochs, loss, t_loss, accuracy))
plt.plot(train_loss)
def test_network_fit():
network = NeuralNetwork(layers=[2, 3, 1])
x = np.array([[2, 3], [1, 1]])
y = np.array([2, 1]).reshape(2, 1)
cost_history = network.fit(x, y, alpha=0.01, iterations=1000)
print(cost_history)