# The error threshold (percent) for each layer of the network.
# After going below the threshold the pre-training will move
# on to the next layer
deeplearn.setErrorThresholds([0.8, 0.8])
# The learning rate in the range 0.0-1.0
deeplearn.setLearningRate(0.2)
# The percentage of dropouts in the range 0-100
deeplearn.setDropoutsPercent(0.001)
# Title of the training error image
deeplearn.setPlotTitle("Xor Training")
# The number of time steps after which the training error image is redrawn
deeplearn.setHistoryPlotInterval(50000)
print "Training started"
timeStep = 0
while (deeplearn.training() != 0):
timeStep = timeStep + 1
print "Training Completed"
print "Test data set performance is " + str(deeplearn.getPerformance()) + "%";
deeplearn.export("result.py")
print "Exported trained network"
deeplearn.save("result.nn")
print "Saved trained network"
# The error threshold (percent) for each layer of the network.
# After going below the threshold the pre-training will move
# on to the next layer
deeplearn.setErrorThresholds([0.8, 0.8])
# The learning rate in the range 0.0-1.0
deeplearn.setLearningRate(0.2)
# The percentage of dropouts in the range 0-100
deeplearn.setDropoutsPercent(0.001)
# Title of the training error image
deeplearn.setPlotTitle("Xor Training")
# The number of time steps after which the training error image is redrawn
deeplearn.setHistoryPlotInterval(50000)
print("Training started")
timeStep = 0
while (deeplearn.training() != 0):
timeStep = timeStep + 1
print("Training Completed")
print("Test data set performance is " + str(deeplearn.getPerformance()) + "%")
deeplearn.export("result.py")
print("Exported trained network")
deeplearn.save("result.nn")
print("Saved trained network")
