1 import numpy as np
2 from keras.datasets import mnist
3 from keras.utils import np_utils
4 from keras.models import Sequential
5 from keras.layers import Dense
6 from keras.optimizers import SGD
7 from keras.models import load_model
1 # 载入数据
2 (x_train,y_train),(x_test,y_test) = mnist.load_data()
3 # (60000,28,28)
4 print('x_shape:',x_train.shape)
5 # (60000)
6 print('y_shape:',y_train.shape)
7 # (60000,28,28)->(60000,784)
8 x_train = x_train.reshape(x_train.shape[0],-1)/255.0
9 x_test = x_test.reshape(x_test.shape[0],-1)/255.0
10 # 换one hot格式
11 y_train = np_utils.to_categorical(y_train,num_classes=10)
12 y_test = np_utils.to_categorical(y_test,num_classes=10)
13
14 # 载入模型
15 model = load_model('model.h5')
16
17 # 评估模型
18 loss,accuracy = model.evaluate(x_test,y_test)
19
20 print('
test loss',loss)
21 print('accuracy',accuracy)
![](https://img2018.cnblogs.com/blog/1776985/201909/1776985-20190922124846547-1067496313.jpg)
# 训练模型
model.fit(x_train,y_train,batch_size=64,epochs=2)
# 评估模型
loss,accuracy = model.evaluate(x_test,y_test)
print('
test loss',loss)
print('accuracy',accuracy)
![](https://img2018.cnblogs.com/blog/1776985/201909/1776985-20190922125046652-1953864818.jpg)
# 保存参数,载入参数
model.save_weights('my_model_weights.h5')
model.load_weights('my_model_weights.h5')
# 保存网络结构,载入网络结构
from keras.models import model_from_json
json_string = model.to_json()
model = model_from_json(json_string)
![](https://img2018.cnblogs.com/blog/1776985/201909/1776985-20190922125350126-1734313327.jpg)