I want to correct the part of the question error from deep learning 1 that I make from scratch.

import sys,os
import numpy as np
sys.path.append(os.pardir)
from dataset.mnist import load_mnist
import pickle
from common.functions import sigmoid, softmax


# Create a function to output test data
def get_data():
    # Read MNIST dataset
    (x_train, t_train), (x_test, t_test) = load_mnist (normalize = True, flatten = True, one_hot_label = False)
    
    # Output test data
    return x_test, t_test

# Implement a function to read learned parameters
default_network():
    
    # Loading Learned Parameters
    with open("."/deep-learning-from-scratch-master/ch03/sample_weight.pkl", 'rb') asf:
        network=pickle.load(f)
        
        # Outputs a dictionary containing learned parameters
        return network

# Implement a function that predicts correct answers from handwritten numbers
def predict (network, x):
    # Retrieving Learned Parameters from Dictionary
    W1, W2, W3 = network ['W1'], network ['W2', network ['W3']
    b1, b2, b3 = network['b1'], network['b2'], network['b3']
    
    # first-tier calculation
    a1 = np.dot(x,W1) + b1# weighted sum
    z1 = sigmoid(a1)# activation
    
    # second-tier calculation
    a2 = np.dot(z1,W2) + b2 # weighted sum
    z2=sigmoid(a2)#activation
    
    # third-tier calculation
    a3=np.dot(z2,W3)+b3# weighted sum
    y=softmax(a3)#activation
    
    # Output inference results (output of neural network)
    returny

   # Get test images and labels
x,t=get_data()

# Get Learned Parameters
network=init_network

acuracy_cnt = 0
for i in range (len(x)):
    y=predict(network, x[i])
    p=np.argmax(y)#Get index of the most likely element
    if p==t[i]:
        acuracy_cnt+=1

print("Accuracy:"+str(float(accuracy_cnt)/len(x)))))

Traceback (most recent call last):
  File "c:\deep-learning-from-scratch-master\3.6.2.py", line 56, in<module>
    y=predict(network, x[i])
  File "c:\deep-learning-from-scratch-master\3.6.2.py", line 30, predict
    W1, W2, W3 = network ['W1'], network ['W2', network ['W3']
TypeError: 'function' object is not subscriptable