def _new_initial(self, size, deterministic, more_replacements=None):
aesara_condition_is_here = isinstance(deterministic, Variable)
if aesara_condition_is_here:
return at.switch(
deterministic,
at.repeat(self.mean.dimshuffle("x", 0),
size if size is not None else 1, -1),
self.histogram[self.randidx(size)],
)
else:
if deterministic:
return at.repeat(self.mean.dimshuffle("x", 0),
size if size is not None else 1, -1)
else:
return self.histogram[self.randidx(size)]
def __call__(self, X):
XY = X.dot(X.T)
x2 = at.sum(X**2, axis=1).dimshuffle(0, "x")
X2e = at.repeat(x2, X.shape[0], axis=1)
H = X2e + X2e.T - 2.0 * XY
V = at.sort(H.flatten())
length = V.shape[0]
# median distance
m = at.switch(
at.eq((length % 2), 0),
# if even vector
at.mean(V[((length // 2) - 1):((length // 2) + 1)]),
# if odd vector
V[length // 2],
)
h = 0.5 * m / at.log(floatX(H.shape[0]) + floatX(1))
# RBF
Kxy = at.exp(-H / h / 2.0)
# Derivative
dxkxy = -at.dot(Kxy, X)
sumkxy = at.sum(Kxy, axis=-1, keepdims=True)
dxkxy = at.add(dxkxy, at.mul(X, sumkxy)) / h
return Kxy, dxkxy
def _new_initial(self, size, deterministic, more_replacements=None):
aesara_condition_is_here = isinstance(deterministic, Variable)
if size is None:
size = 1
size = at.as_tensor(size)
if aesara_condition_is_here:
return at.switch(
deterministic,
at.repeat(self.mean.reshape((1, -1)), size, -1),
self.histogram[self.randidx(size)],
)
else:
if deterministic:
raise NotImplementedInference(
"Deterministic sampling from a Histogram is broken in v4")
return at.repeat(self.mean.reshape((1, -1)), size, -1)
else:
return self.histogram[self.randidx(size)]
def logdet(self):
return aet.repeat(aet.sum(aet.log(self.scale)), self.z0.shape[0])
