python错误解决

Question

class Net:
    def __init__(self,input_size,hidden_size,output_size,weight_init_std=0.001):
        self.params={}
        self.params['W1']=weight_init_std* \
                          np.random.randn(input_size,hidden_size)#高斯分布初始化
        self.params['W2']=weight_init_std* \
                          np.random.randn(hidden_size,output_size)
        self.params['b1']=np.zeros(hidden_size)
        self.params['b2']=np.zeros(output_size)
    def sigmoid(self,x):
        x=np.array(x)
        return 1/(1+np.exp(-x))

    def softmax(self,x):
        c=np.max(x)
        exp_a=np.exp(x-c)#溢出对策
        sum_exp=np.sum(exp_a)
        y=exp_a/sum_exp
        return y
    def predict(self,x):
        W1,W2=self.params['W1'],self.params['W2']
        b1,b2=self.params['b1'],self.params['b2']
        a1=np.dot(x,W1)+b1
        z1=self.sigmoid(a1)
        a2=np.dot(z1,W2)+b2
        z2=self.sigmoid(a2)
        y=self.softmax(z2)
        return y
    def loss(self,x,t):
        y=self.predict(x)
        if y.ndim==1:
            t=t.reshape(1,t.size)
            y=y.reshape(1,y.size)
        batch_size=y.shape[0]
        return -np.sum(t*np.log(y+1e-7))/batch_size
    def accuracy(self,x,t):
        y=self.predict(x)
        y=np.argmax(y,axis=1)
        t=np.argmax(t,axis=1)
        accuracy=np.sum(y==t)/float(x.shape[0])
        return accuracy
    def numerical_gradient(self,f,x):#求偏导
        h=1e-4
        grad=np.zeros_like(x)
        tmp_val=np.zeros_like(x[0])
        for idx in range(x.size):
            tmp_val=x[idx]#初始值
            x[idx]=tmp_val+h
            fxh1=f(x)
            x[idx]=tmp_val-h
            fxh2=f(x)
            grad[idx]=(fxh1-fxh2)/(2*h)
            x[idx]=tmp_val#还原x
        return grad
    def ng(self,x,t):
        f=lambda w:self.loss(x,t)
        grads={}
        grads['W1']=self.numerical_gradient(f,self.params['W1'])
        grads['b1']=self.numerical_gradient(f,self.params['b1'])
        grads['W2']=self.numerical_gradient(f,self.params['W2'])
        grads['b2']=self.numerical_gradient(f,self.params['b2'])
        return grads
net=Net(input_size=784,hidden_size=225,output_size=15)
x=np.random.rand(100,784)
y=net.predict(x)
t=np.random.rand(100,15)
grads=net.ng(x,t)

错误提示：

Answer 1

代码46行

for idx in range(x.size)改for idx in range(x.shape[0]):