def main():
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
myMLP = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.005,
epochs=30)
# myMLP.fortest()
myMLP.train()
# Do the recognizer
# Explicitly specify the test set to be evaluated
MLPPred = myMLP.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("\nResult of the MLP:")
evaluator.printAccuracy(data.testSet, MLPPred)
# Draw
plot = PerformancePlot("MLP validation")
plot.draw_performance_epoch(myMLP.performances, myMLP.epochs)
def runMultilayerClassifier(data):
c = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
epochs=30,
layers=MultilayerPerceptron.createLayers([784, 100, 10], 0.05))
trainAndEvaluateClassifier(c, data.test_set, verbose=True)
def main():
# ------------ NOTE --------------
# oneHot = False, as the framwork provided implements binary one-hot encoding (is it a 7 = True/False)
# Our targets are one-of-k encoded (e.g. 1= (0,1,0,0,0,0,0,0,0)). Network predicts the exact number on the picture not just 7 = True/False
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
myMLPClassifier = MultilayerPerceptron(
data.trainingSet,
data.validationSet,
data.testSet,
hiddenLayerSizes=[
65, 30
], # size of hidden layers, input and output layers sizes are constant
learningRate=0.028, # learning rate
epochs=50) # epochs
# Train the classifiers
print("=========================")
print("Training..")
print("\nMLP has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("Result of the stupid recognizer:")
print("\nResult of the MLP:")
#evaluator.printComparison(data.testSet, lrPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# Draw
plot = PerformancePlot("MLP validation")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)
def classify_all():
data = MNISTSeven("../data/mnist_seven.csv", 4000, 500, 500, one_hot=False)
mlp = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
output_task="classify_all",
cost='crossentropy',
output_activation='softmax',
learning_rate=0.30,
epochs=50)
mlp.train()
pred = mlp.evaluate()
evaluator = Evaluator()
evaluator.printAccuracy(data.test_set, pred)
def classify_all():
data = MNISTSeven("../data/mnist_seven.csv", 4000, 500, 500,
one_hot=False)
mlp = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
output_task="classify_all",
cost='crossentropy',
output_activation='softmax',
learning_rate=0.30,
epochs=50)
mlp.train()
pred = mlp.evaluate()
evaluator = Evaluator()
evaluator.printAccuracy(data.test_set, pred)
def main():
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
myMLPClassifier = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.05,
epochs=20)
#no more changes after 20 epochs
#Removed old stuff
# Report the result #
print("=========================")
evaluator = Evaluator()
# Train the classifiers
print("=========================")
print("Training..")
print("\nMLP has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("\nResult of the MLP recognizer:")
#evaluator.printComparison(data.testSet, lrPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# Draw
plot = PerformancePlot("MLP validation")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)
def main():
# ------------ NOTE --------------
# oneHot does not work: It makes not sense to have binary labeled data (targetDigit or notTargetDigit),
# but having a MLP with 10 output nodes and softmax function. It needs to be trained with digit labels, not binary ones!
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
oneHot=False)
myMLPClassifier = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
hiddenLayerSizes=[65,30], # size of hidden layers, input and output layers sizes are constant
learningRate=0.028, # learning rate
epochs=50) # epochs
# Train the classifiers
print("=========================")
print("Training..")
print("\nMLP has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("Result of the stupid recognizer:")
print("\nResult of the MLP:")
#evaluator.printComparison(data.testSet, lrPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# Draw
plot = PerformancePlot("MLP validation")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)
def main():
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
myMLPClassifier = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
loss='ce',
layers=[128, 150],
learningRate=0.005,
epochs=10)
# Report the result #
# Train the classifiers
print("=========================")
print("Training..")
print("\nMultilayer Perceptron has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("\nResult of the Multi Layer Perceptron recognizer:")
evaluator.printComparison(data.testSet, mlpPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# Draw
plot = PerformancePlot("Multi Layer Perceptron validation")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)
def __init__(self, train, valid, test, learning_rate=0.1, epochs=30):
"""
Parameters
----------
train : list
valid : list
test : list
learning_rate : float
epochs : positive int
Attributes
----------
training_set : list
validation_set : list
test_set : list
learning_rate : float
epochs : positive int
performances: array of floats
"""
self.learning_rate = learning_rate
self.epochs = epochs
self.training_set = train
self.validation_set = valid
self.test_set = test
self.performances = []
self.layers = []
# First hidden layer
number_of_1st_hidden_layer = 400
self.layers.append(LogisticLayer(train.input.shape[1], number_of_1st_hidden_layer, None, activation="tanh", is_classifier_layer=False))
# Output layer
self.layers.append(LogisticLayer(number_of_1st_hidden_layer, train.input.shape[1], None, activation="tanh", is_classifier_layer=True))
self.MLP = MultilayerPerceptron(self.training_set, self.validation_set, self.test_set, layers = self.layers, learning_rate=0.05, epochs=30)
def main():
# data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
one_hot=False)
# NOTE:
# Other 1-digit classifiers do not make sense now for comparison purpose
# So you should comment them out, let alone the MLP training and evaluation
# Train the classifiers #
print("=========================")
print("Training..")
# Stupid Classifier
# myStupidClassifier = StupidRecognizer(data.training_set,
# data.validation_set,
# data.test_set)
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
# Perceptron
# myPerceptronClassifier = Perceptron(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=10)
#
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# perceptronPred = myPerceptronClassifier.evaluate()
# Logistic Regression
# myLRClassifier = LogisticRegression(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=30)
#
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# lrPred = myLRClassifier.evaluate()
# Multi layer perceptron
layers = [LogisticLayer(data.training_set.input.shape[1], 30, None, 'sigmoid', False),
LogisticLayer(30, 10, None, 'softmax', True)]
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
layers=layers,
learning_rate=0.005, epochs=30)
print("\nLogistic Regression has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result #
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
print("\nResult of the Multi-layer Perceptron recognizer (on test set):")
evaluator.printComparison(data.test_set, mlpPred)
evaluator.printAccuracy(data.test_set, mlpPred)
def main():
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
# myLRClassifier = LogisticRegression(data.trainingSet,
# data.validationSet,
# data.testSet,
# learningRate=0.005,
# epochs=30)
hidden_layers = [
LogisticLayer(128, 32, isClassifierLayer=True) for layer in range(1)
]
mlp = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
hidden_layers,
learningRate=0.005,
epochs=30)
# Train the classifiers
#print("=========================")
print("Training...")
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
print("Training MLP...")
mlp.train()
print("Done.")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
# perceptronPred = myPerceptronClassifier.evaluate()
# lrPred = myLRClassifier.evaluate()
mlpPred = mlp.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# # #evaluator.printComparison(data.testSet, stupidPred)
# # evaluator.printAccuracy(data.testSet, stupidPred)
# #
# # print("\nResult of the Perceptron recognizer:")
# # #evaluator.printComparison(data.testSet, perceptronPred)
# # evaluator.printAccuracy(data.testSet, perceptronPred)
# #
# # print("\nResult of the Logistic Regression recognizer:")
# # #evaluator.printComparison(data.testSet, lrPred)
# # evaluator.printAccuracy(data.testSet, lrPred)
print("Result of the MLP recognizer:")
evaluator.printComparison(data.testSet, mlpPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# Draw
plot = PerformancePlot("Logistic Regression validation")
# plot.draw_performance_epoch(myLRClassifier.performances,
# myLRClassifier.epochs)
plot.draw_performance_epoch(mlp.performances, mlp.epochs)
class DenoisingAutoEncoder(AutoEncoder):
"""
A denoising autoencoder.
"""
def __init__(self, train, valid, test, learning_rate=0.1, epochs=30):
"""
Parameters
----------
train : list
valid : list
test : list
learning_rate : float
epochs : positive int
Attributes
----------
training_set : list
validation_set : list
test_set : list
learning_rate : float
epochs : positive int
performances: array of floats
"""
self.learning_rate = learning_rate
self.epochs = epochs
self.training_set = train
self.validation_set = valid
self.test_set = test
self.performances = []
self.layers = []
# First hidden layer
number_of_1st_hidden_layer = 400
self.layers.append(LogisticLayer(train.input.shape[1], number_of_1st_hidden_layer, None, activation="tanh", is_classifier_layer=False))
# Output layer
self.layers.append(LogisticLayer(number_of_1st_hidden_layer, train.input.shape[1], None, activation="tanh", is_classifier_layer=True))
self.MLP = MultilayerPerceptron(self.training_set, self.validation_set, self.test_set, layers = self.layers, learning_rate=0.05, epochs=30)
def train(self, verbose=True):
"""
Train the denoising autoencoder
"""
for epoch in range(self.epochs):
if verbose:
print("Training DAE epoch {0}/{1}.."
.format(epoch + 1, self.epochs))
self._train_one_epoch()
if False:#verbose:
accuracy = accuracy_score(self.validation_set.label,
self.evaluate(self.validation_set))
# Record the performance of each epoch for later usages
# e.g. plotting, reporting..
self.performances.append(accuracy)
print("Accuracy on validation: {0:.2f}%"
.format(accuracy * 100))
print("-----------------------------")
pass
def _train_one_epoch(self):
"""
Train one epoch, seeing all input instances
"""
for img in self.training_set.input:
self.noise = 0.1
noisy = img + self.noise * np.random.uniform(-1.0,1.0)
normalized = Activation.tanh(noisy)
self.MLP._feed_forward(normalized)
self.MLP._compute_error(normalized[1:])
self.MLP._update_weights()
pass
def _get_weights(self):
"""
Get the weights (after training)
"""
return self.MLP._get_input_layer().weights
def main():
#data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
one_hot=False)
# # NOTE:
# # Other 1-digit classifiers do not make sense now for comparison purpose
# # So you should comment them out, let alone the MLP training and evaluation
#
# # Train the classifiers #
# print("=========================")
# print("Training..")
#
# # Stupid Classifier
# myStupidClassifier = StupidRecognizer(data.training_set,
# data.validation_set,
# data.test_set)
#
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
# # Do the recognizer
# # Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
#
# # Perceptron
# myPerceptronClassifier = Perceptron(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=10)
#
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# # Do the recognizer
# # Explicitly specify the test set to be evaluated
# perceptronPred = myPerceptronClassifier.evaluate()
#
# # Logistic Regression
# myLRClassifier = LogisticRegression(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=30)
#
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
# # Do the recognizer
# # Explicitly specify the test set to be evaluated
# lrPred = myLRClassifier.evaluate()
# Build up the network from specific layers
# Here is an example of a MLP acting like the Logistic Regression
layers = []
layers.append(LogisticLayer(784, 5, None, "sigmoid", True))
layers.append(LogisticLayer(5, 10, None, "softmax", False))
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.5,
epochs=30, layers=layers)
print("\nLogistic Regression has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
#
# Report the result #
print("=========================")
evaluator = Evaluator()
#
# # print("Result of the stupid recognizer:")
# # evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# #
# # print("\nResult of the Perceptron recognizer (on test set):")
# # evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# #
# # print("\nResult of the Logistic Regression recognizer (on test set):")
# # evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
#
print("\nResult of the Multi-layer Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
#
# # Draw
# plot = PerformancePlot("Logistic Regression")
# plot.draw_performance_epoch(myLRClassifier.performances,
# myLRClassifier.epochs)
# 3D Plot learning_rates + epochs -> accuracies
print("Creating 3D plot. This may take some minutes...")
learning_rate_sample_count = 5
epochs_sample_count = 20
xticks = np.logspace(-10.0, 0, base=10, num=learning_rate_sample_count, endpoint=False)
accuracies = []
learning_rates = []
epoch_values = []
for i in itertools.product(range(learning_rate_sample_count)):
learning_rate = 100 / np.exp(i)
print("Calculating accuracy for: learning rate = %s" % (learning_rate))
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=learning_rate,
epochs=epochs_sample_count,
layers=layers)
epoch_accuracies = myMLPClassifier.train(False)
lrPred = myMLPClassifier.evaluate()
epoch_values.append([e for e in range(epochs_sample_count)])
learning_rates.append([learning_rate for _ in range(epochs_sample_count)])
accuracies.append(epoch_accuracies)
accuracies_merged = list(itertools.chain(*accuracies))
epochs_merged = list(itertools.chain(*epoch_values))
learning_rates_merged = list(itertools.chain(*learning_rates))
print(accuracies_merged)
print(epochs_merged)
print(learning_rates)
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(np.log10(learning_rates_merged), epochs_merged, accuracies_merged)
ax.set_xlabel("Learning Rate")
ax.set_xticks(np.log10(xticks))
ax.set_xticklabels(xticks)
ax.set_ylabel('Epochs')
ax.set_zlabel('Accuracy')
plt.show()
def main():
# data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
one_hot=False)
# NOTE:
# Other 1-digit classifiers do not make sense now for comparison purpose
# So you should comment them out, let alone the MLP training and evaluation
# Train the classifiers #
print("=========================")
print("Training the autoencoder..")
myDAE = DenoisingAutoEncoder(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.05,
epochs=30)
print("\nAutoencoder has been training..")
myDAE.train()
print("Done..")
# Multi-layer Perceptron
# NOTES:
# Now take the trained weights (layer) from the Autoencoder
# Feed it to be a hidden layer of the MLP, continue training (fine-tuning)
# And do the classification
# Correct the code here
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.05,
epochs=30)
print("\nMulti-layer Perceptron has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result #
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
print("\nResult of the DAE + MLP recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
# Draw
plot = PerformancePlot("DAE + MLP on MNIST task")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)
def main():
# data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
one_hot=False)
# NOTE:
# Other 1-digit classifiers do not make sense now for comparison purpose
# So you should comment them out, let alone the MLP training and evaluation
# Train the classifiers #
print("=========================")
print("Training the autoencoder..")
myDAE = DenoisingAutoEncoder(data.training_set,
data.validation_set,
data.test_set,
learning_rate=dae_lr,
noiseRatio=dae_nr,
hiddenLayerNeurons=hiddenLayerNeurons,
epochs=dae_epochs)
print("\nAutoencoder has been training..")
myDAE.train()
print("Done..")
# Multi-layer Perceptron
# NOTES:
# Now take the trained weights (layer) from the Autoencoder
# Feed it to be a hidden layer of the MLP, continue training (fine-tuning)
# And do the classification
myMLPLayers = []
# First hidden layer
number_of_1st_hidden_layer = hiddenLayerNeurons
myMLPLayers.append(LogisticLayer(data.training_set.input.shape[1]-1, # bias "1" already added so remove one
number_of_1st_hidden_layer,
weights=myDAE._get_weights(),
activation="sigmoid",
is_classifier_layer=False))
# Output layer
number_of_output_layer = 10
myMLPLayers.append(LogisticLayer(number_of_1st_hidden_layer,
number_of_output_layer,
None,
activation="softmax",
is_classifier_layer=True))
# Correct the code here
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
layers=myMLPLayers,
learning_rate=mlp_lr,
epochs=mlp_epochs)
# remove double added bias "1"
myMLPClassifier.__del__()
print("\nMulti-layer Perceptron has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result #
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
print("\nResult of the DAE + MLP recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
os.chdir("..")
# Draw
plot = PerformancePlot("DAE + MLP on MNIST task on validation set")
plot.draw_performance_epoch(myMLPClassifier.performances, myMLPClassifier.epochs, "plots", filename)
print("drawing weights of auto encoder mlp input…")
weight_plotter = WeightVisualizationPlot(myDAE.autoencMLP)
weight_plotter.plot()
def main():
#data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
one_hot=False)
# # NOTE:
# # Other 1-digit classifiers do not make sense now for comparison purpose
# # So you should comment them out, let alone the MLP training and evaluation
#
# # Train the classifiers #
# print("=========================")
# print("Training..")
#
# # Stupid Classifier
# myStupidClassifier = StupidRecognizer(data.training_set,
# data.validation_set,
# data.test_set)
#
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
# # Do the recognizer
# # Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
#
# # Perceptron
# myPerceptronClassifier = Perceptron(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=10)
#
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# # Do the recognizer
# # Explicitly specify the test set to be evaluated
# perceptronPred = myPerceptronClassifier.evaluate()
#
# # Logistic Regression
# myLRClassifier = LogisticRegression(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=30)
#
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
# # Do the recognizer
# # Explicitly specify the test set to be evaluated
# lrPred = myLRClassifier.evaluate()
# Build up the network from specific layers
# Here is an example of a MLP acting like the Logistic Regression
layers = []
layers.append(LogisticLayer(784, 5, None, "sigmoid", True))
layers.append(LogisticLayer(5, 10, None, "softmax", False))
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.5,
epochs=30,
layers=layers)
print("\nLogistic Regression has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
#
# Report the result #
print("=========================")
evaluator = Evaluator()
#
# # print("Result of the stupid recognizer:")
# # evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# #
# # print("\nResult of the Perceptron recognizer (on test set):")
# # evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# #
# # print("\nResult of the Logistic Regression recognizer (on test set):")
# # evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
#
print("\nResult of the Multi-layer Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
#
# # Draw
# plot = PerformancePlot("Logistic Regression")
# plot.draw_performance_epoch(myLRClassifier.performances,
# myLRClassifier.epochs)
# 3D Plot learning_rates + epochs -> accuracies
print("Creating 3D plot. This may take some minutes...")
learning_rate_sample_count = 5
epochs_sample_count = 20
xticks = np.logspace(-10.0,
0,
base=10,
num=learning_rate_sample_count,
endpoint=False)
accuracies = []
learning_rates = []
epoch_values = []
for i in itertools.product(range(learning_rate_sample_count)):
learning_rate = 100 / np.exp(i)
print("Calculating accuracy for: learning rate = %s" % (learning_rate))
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=learning_rate,
epochs=epochs_sample_count,
layers=layers)
epoch_accuracies = myMLPClassifier.train(False)
lrPred = myMLPClassifier.evaluate()
epoch_values.append([e for e in range(epochs_sample_count)])
learning_rates.append(
[learning_rate for _ in range(epochs_sample_count)])
accuracies.append(epoch_accuracies)
accuracies_merged = list(itertools.chain(*accuracies))
epochs_merged = list(itertools.chain(*epoch_values))
learning_rates_merged = list(itertools.chain(*learning_rates))
print(accuracies_merged)
print(epochs_merged)
print(learning_rates)
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(np.log10(learning_rates_merged), epochs_merged,
accuracies_merged)
ax.set_xlabel("Learning Rate")
ax.set_xticks(np.log10(xticks))
ax.set_xticklabels(xticks)
ax.set_ylabel('Epochs')
ax.set_zlabel('Accuracy')
plt.show()
def main():
# data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
one_hot=False)
# NOTE:
# Other 1-digit classifiers do not make sense now for comparison purpose
# So you should comment them out, let alone the MLP training and evaluation
# Train the classifiers #
print("=========================")
print("Training the autoencoder..")
myDAE = DenoisingAutoEncoder(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.05,
epochs=10)
print("\nAutoencoder has been training..")
myDAE.train()
print("Done..")
# Multi-layer Perceptron
# NOTES:
# Now take the trained weights (layer) from the Autoencoder
# Feed it to be a hidden layer of the MLP, continue training (fine-tuning)
# And do the classification
# Correct the code here
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.05,
epochs=30, input_weights=myDAE._get_weights())
print("\nMulti-layer Perceptron has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result #
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
print("\nResult of the DAE + MLP recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
# Draw
#plot = PerformancePlot("DAE + MLP on MNIST task")
#plot.draw_performance_epoch(myMLPClassifier.performances,
# myMLPClassifier.epochs)
#print myDAE._get_weights().shape[1]
weights = 0.5 * myDAE._get_weights() + 0.5
wplot = WeightVisualizationPlot(weights)
wplot.draw_weights()
def main():
data = MNISTSeven("data/mnist_seven.csv", 3000, 1000, 1000, oneHot=True)
myStupidClassifier = StupidRecognizer(data.trainingSet, data.validationSet,
data.testSet)
myPerceptronClassifier = Perceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.005,
epochs=30)
myLRClassifier = LogisticRegression(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.005,
epochs=30)
# Report the result #
print("=========================")
evaluator = Evaluator()
# Train the classifiers
print("=========================")
print("Training..")
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
myMLP = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.01,
epochs=30,
loss="ce",
outputActivation="softmax",
weight_decay=0.1)
print("\nMLP has been training..")
myMLP.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
# perceptronPred = myPerceptronClassifier.evaluate()
# lrPred = myLRClassifier.evaluate()
mlpPred = myMLP.evaluate(data.validationSet)
# # Report the result
# print("=========================")
# evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# #evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.testSet, stupidPred)
# print("\nResult of the Perceptron recognizer:")
# #evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.testSet, perceptronPred)
# print("\nResult of the Logistic Regression recognizer:")
# #evaluator.printComparison(data.testSet, lrPred)
# evaluator.printAccuracy(data.testSet, lrPred)
print("\nResult of the Multilayer Perceptron recognizer:")
#evaluator.printComparison(data.testSet, lrPred)
# evaluator.printAccuracy(data.testSet, mlpPred)
plot = PerformancePlot("MLP validation")
plot.draw_performance_epoch(myMLP.performances, myMLP.epochs)
def main():
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000)
myStupidClassifier = StupidRecognizer(data.trainingSet,
data.validationSet,
data.testSet)
myPerceptronClassifier = Perceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.005,
epochs=30)
myLRClassifier = LogisticRegression(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.005,
epochs=30)
MlpClassifier = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.1,
epochs = 30)
# Train the classifiers
print("=========================")
print("Training..")
print("\nStupid Classifier has been training..")
myStupidClassifier.train()
print("Done..")
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
print("\nStarting Backpropagation MLP training...")
MlpClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
stupidPred = myStupidClassifier.evaluate()
# perceptronPred = myPerceptronClassifier.evaluate()
# lrPred = myLRClassifier.evaluate()
mlpPred = MlpClassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
evaluator.printAccuracy(data.testSet, stupidPred)
# print("\nResult of the Perceptron recognizer:")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.testSet, perceptronPred)
# print("\nResult of the Logistic Regression recognizer:")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.testSet, lrPred)
print("\nResult of the MLP recognizer:")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# eval.printConfusionMatrix(data.testSet, pred)
# eval.printClassificationResult(data.testSet, pred, target_names)
print("=========================")
def main():
# data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
one_hot=False)
# NOTE:
# Other 1-digit classifiers do not make sense now for comparison purpose
# So you should comment them out, let alone the MLP training and evaluation
# Train the classifiers #
print("=========================")
print("Training..")
# Stupid Classifier
# myStupidClassifier = StupidRecognizer(data.training_set,
# data.validation_set,
# data.test_set)
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
# Perceptron
# myPerceptronClassifier = Perceptron(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=10)
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# perceptronPred = myPerceptronClassifier.evaluate()
# Logistic Regression
# myLRClassifier = LogisticRegression(data.training_set,
# data.validation_set,
# data.test_set,
# learning_rate=0.005,
# epochs=30)
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# lrPred = myLRClassifier.evaluate()
# Multi-layer Perceptron
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.05,
epochs=30)
print("\nMulti-layer Perceptron has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result #
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
print("\nResult of the Multi-layer Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
# Draw
plot = PerformancePlot("Multi-layer Perceptron on MNIST task")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)
def classify_one():
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
one_hot=True, target_digit='7')
# NOTE:
# Comment out the MNISTSeven instantiation above and
# uncomment the following to work with full MNIST task
# data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
# one_hot=False)
# NOTE:
# Other 1-digit classifiers do not make sense now for comparison purpose
# So you should comment them out, let alone the MLP training and evaluation
# Train the classifiers #
print("=========================")
print("Training..")
# Stupid Classifier
myStupidClassifier = StupidRecognizer(data.training_set,
data.validation_set,
data.test_set)
print("\nStupid Classifier has been training..")
myStupidClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
stupidPred = myStupidClassifier.evaluate()
# Perceptron
myPerceptronClassifier = Perceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.005,
epochs=10)
print("\nPerceptron has been training..")
myPerceptronClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
perceptronPred = myPerceptronClassifier.evaluate()
# Logistic Regression
myLRClassifier = LogisticRegression(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.20,
epochs=30)
print("\nLogistic Regression has been training..")
myLRClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
lrPred = myLRClassifier.evaluate()
# Logistic Regression
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
learning_rate=0.30,
epochs=50)
print("\nMultilayer Perceptron has been training..")
myMLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
mlpPred = myMLPClassifier.evaluate()
# Report the result #
print("=========================")
evaluator = Evaluator()
print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
evaluator.printAccuracy(data.test_set, stupidPred)
print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, perceptronPred)
print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, lrPred)
print("\nResult of the Multi-layer Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
# Draw
plot = PerformancePlot("Logistic Regression")
plot.draw_performance_epoch(myLRClassifier.performances,
myLRClassifier.epochs)
def main():
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
myStupidClassifier = StupidRecognizer(data.trainingSet, data.validationSet,
data.testSet)
#myPerceptronClassifier = Perceptron(data.trainingSet,
#data.validationSet,
#data.testSet,
#learningRate=0.005,
#epochs=30)
#myLRClassifier = LogisticRegression(data.trainingSet,
#data.validationSet,
#data.testSet,
#learningRate=0.005,
#epochs=30)
mlp = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
layers=None,
inputWeights=None,
outputTask='classification',
outputActivation='softmax',
loss='cee',
learningRate=0.01,
epochs=50)
# Report the result #
print("=========================")
evaluator = Evaluator()
# Train the classifiers
print("=========================")
print("Training..")
print("\nStupid Classifier has been training..")
myStupidClassifier.train()
print("Done..")
print("\nPerceptron has been training..")
#myPerceptronClassifier.train()
print("Done..")
print("\nLogistic Regression has been training..")
#myLRClassifier.train()
print("Done..")
print("\nmlp has been training..")
mlp.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
stupidPred = myStupidClassifier.evaluate()
#perceptronPred = myPerceptronClassifier.evaluate()
#lrPred = myLRClassifier.evaluate()
mlppred = MultilayerPerceptron.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
print("Result of the stupid recognizer:")
#evaluator.printComparison(data.testSet, stupidPred)
evaluator.printAccuracy(data.testSet, stupidPred)
print("\nResult of the Perceptron recognizer:")
#evaluator.printComparison(data.testSet, perceptronPred)
#evaluator.printAccuracy(data.testSet, perceptronPred)
print("\nResult of the Logistic Regression recognizer:")
#evaluator.printComparison(data.testSet, lrPred)
#evaluator.printAccuracy(data.testSet, lrPred)
print("Result of the mlp:")
evaluator.printAccuracy(data.testSet, mlppred)
# Draw
#plot = PerformancePlot("Logistic Regression validation")
#plot.draw_performance_epoch(myLRClassifier.performances,
#myLRClassifier.epochs)
####可能有问题
plot = PerformancePlot("mlp validation")
plot.draw_performance_epoch(mlp.performances, mlp.epochs)
def main():
data = MNISTSeven("../data/mnist_seven.csv",
3000,
1000,
1000,
oneHot=False)
# myStupidClassifier = StupidRecognizer(data.trainingSet,
# data.validationSet,
# data.testSet)
# myPerceptronClassifier = Perceptron(data.trainingSet,
# data.validationSet,
# data.testSet,
# learningRate=0.005,
# epochs=30)
#
# myLRClassifier = LogisticRegression(data.trainingSet,
# data.validationSet,
# data.testSet,
# learningRate=0.005,
# epochs=30)
myMLPlassifier = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
learningRate=0.005,
epochs=30)
# Report the result #
print("=========================")
evaluator = Evaluator()
# Train the classifiers
print("=========================")
print("Training..")
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
#
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
print("\nMLP has been training..")
myMLPlassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
# perceptronPred = myPerceptronClassifier.evaluate()
# lrPred = myLRClassifier.evaluate()
mplPred = myMLPlassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
#evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.testSet, stupidPred)
# print("\nResult of the Perceptron recognizer:")
# #evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.testSet, perceptronPred)
#
# print("\nResult of the Logistic Regression recognizer:")
# #evaluator.printComparison(data.testSet, lrPred)
# evaluator.printAccuracy(data.testSet, lrPred)
print("Result of the MLP recognizer:")
#evaluator.printComparison(data.testSet, stupidPred)
evaluator.printAccuracy(data.testSet, mplPred)
# Draw
plot = PerformancePlot("MLP validation")
plot.draw_performance_epoch(myMLPlassifier.performances,
myMLPlassifier.epochs)
def main():
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
oneHot=False)
data.trainingSet.input = np.insert(data.trainingSet.input, 0, 1,
axis=1)
data.validationSet.input = np.insert(data.validationSet.input, 0, 1,
axis=1)
data.testSet.input = np.insert(data.testSet.input, 0, 1, axis=1)
myStupidClassifier = StupidRecognizer(data.trainingSet,
data.validationSet,
data.testSet)
# myPerceptronClassifier = Perceptron(data.trainingSet,
# data.validationSet,
# data.testSet,
# learningRate=0.005,
# epochs=30)
# myLRClassifier = LogisticRegression(data.trainingSet,
# data.validationSet,
# data.testSet,
# learningRate=0.005,
# epochs=30)
MLPClassifier = MultilayerPerceptron(data.trainingSet,
data.validationSet,
data.testSet,
netStruct = [800, 100, 10],
actFunc = ['relu', 'relu', 'softmax'],
dropout = True,
loss = 'crossentropy',
learningRate = 0.001,
epochs = 300)
# Report the result #
print("=========================")
evaluator = Evaluator()
# Train the classifiers
print("=========================")
print("Training..")
# print("\nStupid Classifier has been training..")
# myStupidClassifier.train()
# print("Done..")
#
# print("\nPerceptron has been training..")
# myPerceptronClassifier.train()
# print("Done..")
# print("\nLogistic Regression has been training..")
# myLRClassifier.train()
# print("Done..")
print("\nMLP has been training..")
MLPClassifier.train()
print("Done..")
# Do the recognizer
# Explicitly specify the test set to be evaluated
# stupidPred = myStupidClassifier.evaluate()
# perceptronPred = myPerceptronClassifier.evaluate()
# lrPred = myLRClassifier.evaluate()
mlpPred = MLPClassifier.evaluate()
# Report the result
print("=========================")
evaluator = Evaluator()
# print("Result of the stupid recognizer:")
# #evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.testSet, stupidPred)
#
# print("\nResult of the Perceptron recognizer:")
# #evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.testSet, perceptronPred)
# print("\nResult of the Logistic Regression recognizer:")
# #evaluator.printComparison(data.testSet, lrPred)
# evaluator.printAccuracy(data.testSet, lrPred)
print("\nResult of the MLP recognizer:")
# evaluator.printComparison(data.testSet, lrPred)
evaluator.printAccuracy(data.testSet, mlpPred)
# Draw
# plot = PerformancePlot("MLP validation")
# plot.draw_performance_epoch(MLPClassifier.performances,
# MLPClassifier.epochs)
plt.plot(range(MLPClassifier.epochs), MLPClassifier.performances, 'r--')
plt.show()
def main():
data = MNISTSeven("../data/mnist_seven.csv", 3000, 1000, 1000,
one_hot=False)
# Train
# Denoising Auto Encoder
n_encoder_neurons = 100
myDAE = DenoisingAutoEncoder(data.training_set,
data.validation_set,
data.test_set,
n_hidden_neurons=n_encoder_neurons,
noise_ratio=0.2,
learning_rate=0.05,
epochs=5)
print("Train autoencoder..")
myDAE.train(verbose=True)
print("Done..")
# Multilayer Perceptron
layers = []
# Add auto envoder hidden layer.
layers.append(LogisticLayer(data.training_set.input.shape[1], n_encoder_neurons, weights=myDAE.get_weights(), cost="mse", activation="sigmoid", learning_rate=0.05))
# Add another hidden layer just like in the previous exercise.
n_second_hidden_neurons = 100
layers.append(LogisticLayer(n_encoder_neurons, n_second_hidden_neurons, cost="mse", activation="sigmoid", learning_rate=0.05))
# Add output classifier layer with one neuron per digit.
n_out_neurons = 10
layers.append(LogisticLayer(n_second_hidden_neurons, n_out_neurons, cost="crossentropy", activation="softmax", learning_rate=0.05))
myMLPClassifier = MultilayerPerceptron(data.training_set,
data.validation_set,
data.test_set,
layers=layers,
epochs=15)
print("Train MLP..")
myMLPClassifier.train(verbose=True)
print("Done..")
print("")
# Evaluate
print("Evaluate..")
mlpPred = myMLPClassifier.evaluate()
print("Done..")
print("")
print("Results:")
evaluator = Evaluator()
print("")
# print("Result of the stupid recognizer:")
# evaluator.printComparison(data.testSet, stupidPred)
# evaluator.printAccuracy(data.test_set, stupidPred)
# print("\nResult of the Perceptron recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, perceptronPred)
# print("\nResult of the Logistic Regression recognizer (on test set):")
# evaluator.printComparison(data.testSet, perceptronPred)
# evaluator.printAccuracy(data.test_set, lrPred)
# evaluator.printComparison(data.testSet, perceptronPred)
evaluator.printAccuracy(data.test_set, mlpPred)
# Draw
plot = PerformancePlot("DAE + MLP on MNIST task")
plot.draw_performance_epoch(myMLPClassifier.performances,
myMLPClassifier.epochs)