def _mean_visibles_theano(self, h, v):
"""
Computes the probabilities P(v=1|h).
Parameters
----------
h : array-like, shape (n_samples, n_groups, n_components)
values of the hidden layer.
v : The input original Visible Nodes
Returns
-------
v: array-like,shape (n_samples, n_features)
"""
activations = np.array([
convTheano(h[:, i, :], self.components_[i], border='full') +
self.intercept_visible_ for i in range(self.n_groups)
]).sum(axis=0)
visibles = np.array(v)
windowSize = self.window_size
visualSize = int(sqrt(v.shape[1]))
innerSize = visualSize - 2 * windowSize + 2
n_sample = v.shape[0]
innerV = logistic_sigmoid(activations)
innerV = innerV.reshape(
n_sample, visualSize,
visualSize)[:, windowSize - 1:visualSize - windowSize + 1,
windowSize - 1:visualSize - windowSize + 1]
visibles = visibles.reshape(n_sample, visualSize, visualSize)
visibles[:, windowSize - 1:visualSize - windowSize + 1,
windowSize - 1:visualSize - windowSize + 1] = innerV
visibles = visibles.reshape(n_sample, -1)
return visibles
def testConvTheanoMulti():
visibleSample = 3
visibleNodes = np.ones((3, 28 * 28))
h = np.arange(5 * 24 * 24)
h = h.reshape(5, 24 * 24)
h = np.array((h, 2 * h, 3 * h))
current = time.time()
for i in range(1000):
a = np.array(
[convTheano(visibleNodes[i, :], h[i, :, :]) for i in range(3)])
print time.time() - current
current = time.time()
for i in range(1000):
b = convTheano(visibleNodes, h.reshape(3 * 5, 24 * 24))
print time.time() - current
return a, b
def _mean_hiddens_theano(self, v):
"""Computes the probabilities P(h=1|v).
Parameters
----------
v : array-like, shape (n_samples, n_features)
Values of the visible layer.
Returns
-------
h : array-like, shape (n_samples, n_groups, n_components)
Corresponding mean field values for the hidden layer.
"""
activationsWithoutIntercept = convTheano(v, self.components_)
activations = np.array([
activationsWithoutIntercept[:, i, :] + self.intercept_hidden_[i]
for i in range(self.n_groups)
])
n_samples = v.shape[0]
return logistic_sigmoid(
activations.reshape(n_samples * self.n_groups,
self.n_components)).reshape(
n_samples, self.n_groups,
self.n_components)
def _mean_visibles_theano(self,h,v):
"""
Computes the probabilities P(v=1|h).
Parameters
----------
h : array-like, shape (n_samples, n_groups, n_components)
values of the hidden layer.
v : The input original Visible Nodes
Returns
-------
v: array-like,shape (n_samples, n_features)
"""
activations = np.array([convTheano(h[:,i,:],self.components_[i],border='full') + self.intercept_visible_ for i in range(self.n_groups)]).sum(axis = 0)
visibles = np.array(v)
windowSize = self.window_size
visualSize = int(sqrt(v.shape[1]))
innerSize = visualSize - 2 * windowSize + 2
n_sample = v.shape[0]
innerV = logistic_sigmoid(activations)
innerV = innerV.reshape(n_sample,visualSize, visualSize)[:,windowSize - 1:visualSize - windowSize + 1, windowSize - 1: visualSize - windowSize + 1]
visibles = visibles.reshape(n_sample,visualSize,visualSize)
visibles[:,windowSize - 1: visualSize - windowSize + 1,windowSize - 1: visualSize - windowSize + 1] = innerV
visibles = visibles.reshape(n_sample, -1)
return visibles
def testConvTheano(border='valid'):
a = np.array((1, 0, 1, 0, 1, 0))
b = np.array((0, 1, 0, 1, 0, 1))
A = np.array((a, b, a, b, a, b))
B = np.array((b, a, b, a, b, a))
c = np.array((1, 1, 1))
C = np.array((c, c, c))
D = np.zeros((3, 3))
D[1, 1] = 1
Z = np.array((A, B))
#print Z.shape
Y = np.array((C, D))
#print Y.shape
return convTheano(Z.reshape(2, -1), Y.reshape(2, -1))
def _mean_hiddens_theano(self,v):
"""Computes the probabilities P(h=1|v).
Parameters
----------
v : array-like, shape (n_samples, n_features)
Values of the visible layer.
Returns
-------
h : array-like, shape (n_samples, n_groups, n_components)
Corresponding mean field values for the hidden layer.
"""
activationsWithoutIntercept = convTheano(v,self.components_)
activations = np.array([activationsWithoutIntercept[:,i,:] + self.intercept_hidden_[i] for i in range(self.n_groups)])
n_samples = v.shape[0]
return logistic_sigmoid(activations.reshape(n_samples * self.n_groups, self.n_components)).reshape(n_samples,self.n_groups,self.n_components)
def _gradience_theano(self,v,mean_h):
"""Computer the gradience given the v and h.
This is for getting the Grad0k./ If it is, we need to focus on the Ph0k
Parameters
----------
v: array-like, shape (n_samples, n_features)
values of the visible layer.
h: array-like, shape (n_samples, n_groups, n_components)
Returns
--------
Grad: array-like,shape (n_groups, weight_windowSize * weight_windowSize)
"""
#weights = np.array([convTheano(v[i,:],mean_h[i,:,:]) for i in range(v.shape[0])]).sum(axis = 0)
tempWeights = convTheano(v,mean_h.reshape(v.shape[0] * self.n_groups,self.n_components)).reshape(v.shape[0],v.shape[0],self.n_groups,-1)
#tempWeights = np.array([tempWeights[i,i,:,:] for i in range(v.shape[0])])
tempWeights = tempWeights[0]
weights = tempWeights.sum(axis = 0)
return weights
def _gradience_theano(self, v, mean_h):
"""Computer the gradience given the v and h.
This is for getting the Grad0k./ If it is, we need to focus on the Ph0k
Parameters
----------
v: array-like, shape (n_samples, n_features)
values of the visible layer.
h: array-like, shape (n_samples, n_groups, n_components)
Returns
--------
Grad: array-like,shape (n_groups, weight_windowSize * weight_windowSize)
"""
#weights = np.array([convTheano(v[i,:],mean_h[i,:,:]) for i in range(v.shape[0])]).sum(axis = 0)
tempWeights = convTheano(
v,
mean_h.reshape(v.shape[0] * self.n_groups,
self.n_components)).reshape(v.shape[0], v.shape[0],
self.n_groups, -1)
#tempWeights = np.array([tempWeights[i,i,:,:] for i in range(v.shape[0])])
tempWeights = tempWeights[0]
weights = tempWeights.sum(axis=0)
return weights
