[Microsoft Cognitive Toolkit]Classify Cancer Using Feed Forward Network

1. # Compute the moving average loss to smooth out the noise in SGD
   
2. plotdata["avgloss"] = moving_average(plotdata["loss"])
   
3. plotdata["avgerror"] = moving_average(plotdata["error"])
   
4. 
   
5. # Plot the training loss and the training error
   
6. import matplotlib.pyplot as plt
   
7. 
   
8. plt.figure(1)
   
9. plt.subplot(211)
   
10. plt.plot(plotdata["batchsize"], plotdata["avgloss"], 'b--')
    
11. plt.xlabel('Minibatch number')
    
12. plt.ylabel('Loss')
    
13. plt.title('Minibatch run vs. Training loss')
    
14. 
    
15. plt.show()
    
16. 
    
17. plt.subplot(212)
    
18. plt.plot(plotdata["batchsize"], plotdata["avgerror"], 'r--')
    
19. plt.xlabel('Minibatch number')
    
20. plt.ylabel('Label Prediction Error')
    
21. plt.title('Minibatch run vs. Label Prediction Error')
    
22. plt.show()

复制代码

谢谢楼主分享！

谢谢分享