无法在Jupyter Notebook中调试值错误

时间:2019-07-14 16:56:29

标签: python-3.x deep-learning jupyter-notebook

我有一些ipython笔记本和另一个在此笔记本中使用的辅助python文件“ init_utils.py”。 我得到以下回溯:

ValueError                                Traceback (most recent 
call last)
<ipython-input-42-baebd5bdcbf9> in <module>
----> 1 parameters = model(train_X, train_Y, initialization = 
    "random")
      2 
      3 print("On the train set:")
      4 predictions_train = predict(train_X, train_Y, parameters)
      5 print("On the test set:")

<ipython-input-41-dc37f4190760> in model(X, Y, learning_rate, 
num_iterations, print_cost, initialization)
     34         # Forward propagation: LINEAR -> RELU -> LINEAR ->       
RELU -> LINEAR -> SIGMOID.
     35         print("W2 =",parameters['W2'])
---> 36         a3, cache = forward_propagation(X, parameters)
     37 
     38         # Loss

~/Documents/practice/courses/Deep-Learning-Coursera/Improving Deep 
Neural Networks Hyperparameter tuning, Regularization and 
Optimization/init_utils.py in forward_propagation(X, parameters)    
     62     z1 = np.dot(W1, X) + b1 
     63     a1 = relu(z1)
---> 64     z2 = np.dot(W2, a1) + b2  
     65     a2 = relu(z2)  
     66     z3 = np.dot(W3, a2) + b3

ValueError: shapes (10,2) and (10,300) not aligned: 2 (dim 1) != 
10 (dim 0)

我试图打印W1矩阵,这是它的样子

W2 = [[ 0.74505627  1.97611078]
 [-1.24412333 -0.62641691]
 [-0.80376609 -2.41908317]
 [-0.92379202 -1.02387576]
 [ 1.12397796 -0.13191423]
 [-1.62328545  0.64667545]
 [-0.35627076 -1.74314104]
 [-0.59664964 -0.58859438]
 [-0.8738823   0.02971382]
 [-2.24825777 -0.26776186]]

由于错误,我无法打印a1矩阵

我在这里仅共享相关代码

parameters = model(train_X, train_Y, initialization = "random")
def model(X, Y, learning_rate=0.01, num_iterations=15000, print_cost=True, initialization="he"):
    .
    .
    .
    a3, cache = forward_propagation(X, parameters)
    .
    .
def forward_propagation(X, parameters):
    .
    .
    .
    W2 = parameters["W2"]
    z1 = np.dot(W1, X) + b1
    a1 = relu(z1)
    z2 = np.dot(W2, a1) + b2
    a2 = relu(z2)
    z3 = np.dot(W3, a2) + b3
    a3 = sigmoid(z3)
    .
    .
    .

我无法调试它。谢谢

0 个答案:

没有答案