Question

我正在创建一个将图像发送到服务器的应用程序，服务器将通过神经网络运行图像，并输出结果。这是我正在使用的神经网络类：

 class Network(object):

def __init__(self, sizes):
    self.num_layer = len(sizes)
    self.sizes = sizes
    self.biases = [np.random.randn(y,1) for y in sizes [1:]]
    self.weights = [np.random.randn(y,x) for x, y in zip(sizes[:-1], sizes[1:])]

def feedforward(self, a):
    for b, w in zip(self.biases, self.weights):
        a = sigmoid(np.dot(w, a) + b)
    return a 

def SGD(self, training_data, epochs, mini_batch_size, eta, test_data = None):

    list_training_data = list(training_data)
    if test_data: 
        list_test_Data = list(test_data)
        n_test = len(list_test_Data)
    n = len(list_training_data)
    for j in range(epochs):
        random.shuffle(list_training_data)
        mini_batches = [list_training_data[k:k+mini_batch_size]for k in range(0,n,mini_batch_size)]
        for mini_batch in mini_batches:
            self.update_mini_batch(mini_batch,eta)
        if test_data:
            evulate = self.evaluate(list_test_Data);
            print("Epoch{0}:{1}/{2}" .format( j, evulate, n_test))
        else:
            print("Epoch {0} complete". format(j))

def backdrop(self, x, y):
    nabla_b = [np.zeros(b.shape) for b in self.biases]
    nabla_w = [np.zeros(w.shape) for w in self.weights]
    activation = x
    activations = [x]
    zs = []
    for b, w in zip(self.biases, self.weights):
        z = np.dot(w, activation) + b
        zs.append(z)
        activation = sigmoid(z)
        activations.append(activation)
    delta = self.cost_derivative(activations[-1], y) * \
        sigmoid_prime(zs[-1])
    nabla_b[-1] = delta
    nabla_w[-1] = np.dot(delta, activations[-1-1].transpose())
    return (nabla_b, nabla_w)

def update_mini_batch(self, mini_batch, eta):
    nabla_b = [np.zeros(b.shape) for b in self.biases]
    nabla_w = [np.zeros(w.shape) for w in self.weights]
    for x, y in mini_batch:
        delta_nabla_b, delta_nabla_w = self.backdrop(x, y)
        nabla_b = [nb+dnb for nb, dnb in zip(nabla_b, delta_nabla_b)]
        nabla_w = [nw+dnw for nw, dnw in zip(nabla_w, delta_nabla_w)]
    self.weights = [w-(eta/len(mini_batch))*nw
                    for w ,nw in zip(self.weights, nabla_w)]
    self.biases = [b - (eta/len(mini_batch))*nb
                    for b, nb in zip(self.biases, nabla_b)]

def evaluate(self, test_data):
    test_results = [(np.argmax(self.feedforward(x)), y)for (x, y) in test_data]
    final = sum(int(x==y)for (x,y) in test_results)
    return final

def cost_derivative(self, output_activatoins, y):
    return (output_activatoins - y)

def sigmoid(z):
    return 1.0/(1.0 + np.exp(-z))

def sigmoid_prime(z):
    return sigmoid(z)*(1-sigmoid(z))

我从numpy得到这个错误，我不知道如何解决这个问题。

 C:/Users/name/Desktop/server.py:91: RuntimeWarning: underflow encountered in exp
 np.getter()
 ERROR:__main__:Exception on / [POST]
 Traceback (most recent call last):
 File "C:\Users\name\Downloads\WinPython-64bit-3.4.3.7\python-3.4.3.amd64\lib\site-packages\flask\app.py", line 1817, in wsgi_app
 response = self.full_dispatch_request()
 File "C:\Users\name\Downloads\WinPython-64bit-3.4.3.7\python-3.4.3.amd64\lib\site-packages\flask\app.py", line 1478, in full_dispatch_request
 response = self.make_response(rv)
 File "C:\Users\name\Downloads\WinPython-64bit-3.4.3.7\python-3.4.3.amd64\lib\site-packages\flask\app.py", line 1577, in make_response
rv = self.response_class.force_type(rv, request.environ)
 File "C:\Users\name\Downloads\WinPython-64bit-3.4.3.7\python-3.4.3.amd64\lib\site-packages\werkzeug\wrappers.py", line 847, in force_type
 response = BaseResponse(*_run_wsgi_app(response, environ))
 File "C:\Users\name\Downloads\WinPython-64bit-3.4.3.7\python-3.4.3.amd64\lib\site-packages\werkzeug\test.py", line 871, in run_wsgi_app
app_rv = app(environ, start_response)
TypeError: 'numpy.int64' object is not callable

这是发生错误的行：

def sigmoid(z):
    np.seterr(over='ignore')
    return 1.0/(1.0 + np.exp(-z))

奇怪的是，当我运行神经网络时，我得到了正确的结果。当我通过服务器运行时，我得到TypeError: 'numpy.int64' object is not callable

Answer 1

防止溢出

使用此Sigmoid实现来防止sigmoid上溢和下溢。

def sigmoid_function( signal, derivative=False ):
    # Prevent overflow.
    signal = np.clip( signal, -500, 500 )

    # Calculate activation signal
    signal = 1.0 / (1 + np.exp( -signal ))

    if derivative:
        # Return the partial derivation of the activation function
        return np.multiply(signal, 1-signal)
    else:
        # Return the activation signal
        return signal
#end activation function

输入错误：

我们无法帮助您解决此错误，因为您尚未发布完整代码。但是，当您输入错字并尝试将数字变量作为函数调用时，通常会发生该错误。最后一个问题更适合Code Review StackExchange。

Answer 2

事实证明代码没有任何问题。我收到此错误消息的原因是因为服务器没有以JSON类型发送结果。为了解决这个错误，我从Flask导入了jsonify，并写了return jsonify(results = answer)

使用numpy时出现下溢错误

2 个答案:

防止溢出

输入错误：