Neural networks and deeplearning——錯題簡單分析

這裏給了官網的簡單分析，時間有限我貼圖說道。

Week1

Week2

這題：Convert the entries of a into 0 (if activation <= 0.5) or 1 (if activation > 0.5), stores the predictions in a vector Y_prediction. If you wish, you can use an if/else statement in a for loop (though there is also a way to vectorize this). 我的答案如下：

    for i in range(A.shape[1]):
        
        # Convert probabilities A[0,i] to actual predictions p[0,i]
        ### START CODE HERE ### (≈ 4 lines of code)
        if A[0,i] > 0.5:
            Y_prediction[0,i] = 1
        else: 
            Y_prediction[0,i] = 0
        ### END CODE HERE ###

我不是很熟悉python的技巧，看了一下網友的答案，原來可以一句代碼搞定。

    for i in range(A.shape[1]):
        # Convert probabilities a[0,i] to actual predictions p[0,i]
        ### START CODE HERE ### (≈ 4 lines of code)
        Y_prediction[0, i] = 1 if A[0, i] > 0.5 else 0
        ### END CODE HERE ###

嘻嘻嘻。

圖片太大了可能會失敗。

ValueError: total size of new array must be unchanged

Week3

n_x = X.shape[0]  # size of input layer
n_h = 4
n_y = Y.shape[0]  # size of output layer

initialize_parameters(n_x, n_h, n_y)
A2, cache = forward_propagation(X_assess, parameters)
compute_cost(A2, Y_assess, parameters)
grads = backward_propagation(parameters, cache, X_assess, Y_assess)
parameters = update_parameters(parameters, grads)

predictions = (A2 > 0.5) 

Week4

這題我以爲是激活函數，選錯了。g(z)纔是激活函數。

不要把輸入層算入，最後一層輸出層不是隱藏層。所以L=4，hidden layer = 3

Certificate