1. import numpy as np
2. import nnet as net
3. import matplotlib.pyplot as plt
4. x = np.array([[-1.0, -0.9, -0.8, -0.7, -0.6, -0.5]])
5.  
6. y_t = np.array([[-0.9602, -0.5770, -0.0729, 0.3771, 0.6405, 0.6600]])
7. L = x.shape[0]
8. K1 = y_t.shape[0]
9. w1, b1 = net.nwlog(K1, L)
10. max_epoch = 20000
11. err_goal = 0.01
12. disp_freq = 1000
13. lr = 0.1
14. SSE_vec = []
15. for epoch in range(1, max_epoch+1):
16.     y1 = net.logsig(np.dot(w1, x), b1)
17.     e = y_t - y1
18.     SSE = net.sumsqr(e)
19.     if np.isnan(SSE):
20.         break
21.     SSE_vec.append(SSE)
22.     if SSE < err_goal:
23.         break
24.     d1 = net.deltalog(y1, e)
25.     dw1, db1 = net.learnbp(x, d1, lr)
26.     w1 += dw1
27.     b1 += db1
28.     if (epoch % disp_freq) == 0:
29.         plt.clf()
30.         print("Epoch: ] | SSE: %5.5f " % (epoch, SSE))
31.         plt.plot(x[0], y_t[0], 'r', x[0], y1[0], 'g')
32.         plt.grid()
33.         plt.show()
34.         plt.pause(1e-2)
35.  
36. plt.figure()
37. plt.plot(SSE_vec)
38. plt.ylabel('SSE')
39. plt.yscale('linear')
40. plt.title('epoch')
41. plt.grid(True)
42. plt.show()