# normalising by neuron and by sample nb
finalError = sum(finalError) / (inputSize * testDatasetSize)
print "Final error", finalError
count = 0
for pat in patterns:
print count, pat
count += 1
while True:
ptT = raw_input('Which pattern to test ? (Or Q to quit) : ')
print type(ptT)
if ptT == 'Q' or ptT == 'q':
break
else :
if ptT < str(len(patterns)):
pat = int(ptT)
#print patterns[pat]
sample = [patterns[pat][elem]+0.25*np.random.rand() for elem in range(inputSize)]
print sample
nn.inputData(sample)
nn.computeOutput()
print nn.outputLayer_f
plt.clf()
plt.subplot(2,1,1)
plt.plot(nn.inputLayer)
plt.subplot(2,1,2)
plt.plot(nn.outputLayer_f)
nn.saveWeights()
