#因为训练过程有随机因素,所以读者训练的结果也许会有差别,不必纠结
Trainon60000samples,validateon10000samples
Epoch112
6000060000[==============================]-131s-loss:0。3303-acc:0。8998-val_loss:0。0758-val_acc:0。9766
Epoch212
6000060000[==============================]-9s-loss:
0。1106-acc:0。9676-val_loss:0。0522-val_acc:0。9825
Epoch312
6000060000[==============================]-9s-loss:
0。0831-acc:0。9746-val_loss:0。0405-val_acc:0。9870
Epoch412
6000060000[==============================]-9s-loss:0。0677-acc:0。9798-val_loss:0。0360-val_acc:0。9873
Epoch512
6000060000[==============================]-9s-loss:0。0595-acc:0。9821-val_loss:0。0353-val_acc:0。9875
Epoch612
6000060000[==============================]-10s-loss:0。0556-acc:0。9837-val_loss:0。0327-val_acc:0。9891
Epoch712
6000060000[==============================]-12s-loss:0。0481-acc:0。9855-val_loss:0。0292-val_acc:0。9896
Epoch812
6000060000[==============================]-14s-loss:0。0453-acc:0。9860-val_loss:0。0299-val_acc:0。9897
Epoch912
6000060000[==============================]-17s-loss:0。0420-acc:0。9868-val_loss:0。0297-val_acc:0。9899
Epoch1012
6000060000[==============================]-17s-loss:0。0395-acc:0。9878-val_loss:0。0289-val_acc:0。9904
Epoch1112
6000060000[==============================]-17s-loss:0。0376-acc:0。9885-val_loss:0。0278-val_acc:0。9914
Epoch1212
6000060000[==============================]-15s-loss:0。0357-acc:0。9885-val_loss:0。0268-val_acc:0。9909
#训练结果如下。同样,不同的配置和训练过程会导致总损失不同
#读者实践过程中获得的数值一般不会与教材的示例完全一致
sodel。evaluate(x_test,y_test,verbose=0)
print('Testloss:',score[0])
#使用总损失对模型进行评估,即评估该模型的效果,通过总损失表示
在本教材使用的计算机上,该训练样本的损失如下。
&loss:0。0268492490485