def _get_sample(self, number_samples, resample=False, observed=False, input_values={}, differentiable=True):
if self in input_values:
value = input_values[self]
else:
value = self.value
if not is_discrete(value):
return {self: tile_parameter(value, number_samples=number_samples)}
else:
return {self: value}
def _get_sample(self,
number_samples,
resample=False,
observed=False,
input_values={}):
if self in input_values:
value = input_values[self]
else:
value = self.value
if not is_discrete(value):
return {self: tile_parameter(value, number_samples=number_samples)}
else:
return {
self: value
} #TODO: This is for allowing discrete data, temporary? (for Julia)
def _get_sample(self,
number_samples,
resample=False,
observed=False,
input_values={}):
if self in input_values:
value = input_values[self]
else:
value = self.value
if isinstance(value, chainer.Variable):
return {self: tile_parameter(value, number_samples=number_samples)}
#value_shape = value.shape
#reps = tuple([number_samples] + [1]*len(value_shape[1:]))
#return {self: F.tile(value, reps=reps)} #TODO: Work in progress
else:
return {
self: value
} #TODO: This is for allowing discrete data, temporary?
xc = F.expand_dims(x, axis=dim)
equal_tensor_variable(xt, xc)
##
tensor_tuple = (torch.randn(10, 5), torch.randn(40, 5))
outt = utilities.uniform_shapes(tensor_tuple[0], tensor_tuple[1])
outc = utilities.uniform_shapes_chainer(
chainer.Variable(tensor_tuple[0].numpy()),
chainer.Variable(tensor_tuple[1].numpy()))
all([equal_tensor_variable(xt, xc) for xt, xc in zip(outt, outc)])
## torch.repeat for tiling
ns = 20
x = np.random.normal(size=(1, 20))
xt = utilities.tile_parameter(torch.tensor(x), number_samples=ns)
xc = utilities.tile_parameter_chainer(chainer.Variable(x), number_samples=ns)
equal_tensor_variable(xt, xc)
## get_diagonal: torch reshape
x = np.random.normal(size=(1, 20, 10, 20))
xt = utilities.get_diagonal(torch.tensor(x))
xc = utilities.get_diagonal_chainer(chainer.Variable(x))
equal_tensor_variable(xt, xc)
## sum_from_dim: torch sum
x = np.random.normal(size=(20, 5, 4, 2))
dim_index = 1
xt = utilities.sum_from_dim(torch.tensor(x), dim_index)