def EncodeLabel(self, label):
"""
Arguments:
---------
label: a number between 0 and 9
Returns:
---------
a list of length 10 representing the distributed
encoding of the output.
Description:
-----------
Computes the distributed encoding of a given label.
Example:
-------
0 => [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
3 => [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Notes:
----
Make sure that the elements of the encoding are floats.
"""
target = Target()
target.values = [1.0 if i == label else 0.0 for i in xrange(10)]
return target
def EncodeLabel(self, label):
"""
Arguments:
---------
label: a number between 0 and 9
Returns:
---------
a list of length 10 representing the distributed
encoding of the output.
Description:
-----------
Computes the distributed encoding of a given label.
Example:
-------
0 => [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
3 => [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Notes:
----
Make sure that the elements of the encoding are floats.
"""
# Code seems to expect a Target instance rather than a simple list
# encoded_label = [0.0] * 10
# encoded_label[label] = 1.0
# return encoded_label
new_target = Target()
new_target.values = [0.0] * 10
new_target.values[label] = 1.0
return new_target
def EncodeLabel(self, label):
"""
Arguments:
---------
label: a number between 0 and 9
Returns:
---------
a list of length 10 representing the distributed
encoding of the output.
Description:
-----------
Computes the distributed encoding of a given label.
Example:
-------
0 => [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
3 => [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Notes:
----
Make sure that the elements of the encoding are floats.
"""
List = [0.0 for x in range(10)]
List[label] = 1.0
target = Target()
target.values = List
return target
def Train(network, inputs, targets, learning_rate, epochs):
"""
Arguments:
---------
network : a NeuralNetwork instance
inputs : a list of Input instances
targets : a list of Target instances
learning_rate : a learning_rate (a float)
epochs : a number of epochs (an integer)
Returns:
-------
Nothing
Description:
-----------
This function should train the network for a given number of epochs. That is,
run the *Backprop* over the training set *epochs*-times
"""
network.CheckComplete()
# check if inputs and targets lists have the same length
assert len(inputs) == len(targets)
# run *epoches*-times
for i in range(epochs):
for j in range(len(inputs)):
target = Target()
target.values = targets[j]
Backprop(network, inputs[j], target, learning_rate)
def EncodeLabel(self, label):
"""
Arguments:
---------
label: a number between 0 and 9
Returns:
---------
a list of length 10 representing the distributed
encoding of the output.
Description:
-----------
Computes the distributed encoding of a given label.
Example:
-------
0 => [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
3 => [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Notes:
----
Make sure that the elements of the encoding are floats.
"""
# Code seems to expect a Target instance rather than a simple list
# encoded_label = [0.0] * 10
# encoded_label[label] = 1.0
# return encoded_label
new_target = Target()
new_target.values = [0.0] * 2
new_target.values[label] = 1.0
return new_target
def EncodeLabel(self, label):
"""
Arguments:
---------
// label: a number between 0 and 9
label: a number between 0 and 1
Returns:
---------
// a list of length 10 representing the distributed
encoding of the output.
a list of length 2 representing the distributed
encoding of the output
Description:
-----------
Computes the distributed encoding of a given label.
Example:
-------
0 => [1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
3 => [0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Notes:
----
Make sure that the elements of the encoding are floats.
"""
t = Target()
t.values = [0. for i in range(4)]
t.values[label] = 1.
# if label == 1:
# pick = random.choice([0, 1, 2])
# t.values[pick] = 1.
# else:
# t.values[3] = 1.
# print t.values
return t
