def test_with_some_dynamic_shapes_works(self):
x = array_ops.ones((2, 1, 3))
y = array_ops.placeholder(x.dtype)
z = array_ops.ones(())
with self.test_session() as sess:
bcast_shape = sess.run(
distribution_util.get_broadcast_shape(x, y, z),
feed_dict={y: np.ones((1, 5, 3)).astype(np.float32)})
self.assertAllEqual([2, 5, 3], bcast_shape)
def test_with_some_dynamic_shapes_works(self):
x = array_ops.ones((2, 1, 3))
y = array_ops.placeholder(x.dtype)
z = array_ops.ones(())
with self.test_session() as sess:
bcast_shape = sess.run(
distribution_util.get_broadcast_shape(x, y, z),
feed_dict={y: np.ones((1, 5, 3)).astype(np.float32)})
self.assertAllEqual([2, 5, 3], bcast_shape)
def __init__(self,
concentration1=None,
concentration0=None,
validate_args=False,
allow_nan_stats=True,
name="Kumaraswamy"):
"""Initialize a batch of Kumaraswamy distributions.
Args:
concentration1: Positive floating-point `Tensor` indicating mean
number of successes; aka "alpha". Implies `self.dtype` and
`self.batch_shape`, i.e.,
`concentration1.shape = [N1, N2, ..., Nm] = self.batch_shape`.
concentration0: Positive floating-point `Tensor` indicating mean
number of failures; aka "beta". Otherwise has same semantics as
`concentration1`.
validate_args: Python `bool`, default `False`. When `True` distribution
parameters are checked for validity despite possibly degrading runtime
performance. When `False` invalid inputs may silently render incorrect
outputs.
allow_nan_stats: Python `bool`, default `True`. When `True`, statistics
(e.g., mean, mode, variance) use the value "`NaN`" to indicate the
result is undefined. When `False`, an exception is raised if one or
more of the statistic's batch members are undefined.
name: Python `str` name prefixed to Ops created by this class.
"""
with ops.name_scope(name, values=[concentration1,
concentration0]) as name:
concentration1 = ops.convert_to_tensor(concentration1,
name="concentration1")
concentration0 = ops.convert_to_tensor(concentration0,
name="concentration0")
super(Kumaraswamy, self).__init__(
distribution=uniform.Uniform(
low=array_ops.zeros([], dtype=concentration1.dtype),
high=array_ops.ones([], dtype=concentration1.dtype),
allow_nan_stats=allow_nan_stats),
bijector=bijectors.Kumaraswamy(concentration1=concentration1,
concentration0=concentration0,
validate_args=validate_args),
batch_shape=distribution_util.get_broadcast_shape(
concentration1, concentration0),
name=name)
self._reparameterization_type = distribution.FULLY_REPARAMETERIZED
def __init__(self,
concentration1=None,
concentration0=None,
validate_args=False,
allow_nan_stats=True,
name="Kumaraswamy"):
"""Initialize a batch of Kumaraswamy distributions.
Args:
concentration1: Positive floating-point `Tensor` indicating mean
number of successes; aka "alpha". Implies `self.dtype` and
`self.batch_shape`, i.e.,
`concentration1.shape = [N1, N2, ..., Nm] = self.batch_shape`.
concentration0: Positive floating-point `Tensor` indicating mean
number of failures; aka "beta". Otherwise has same semantics as
`concentration1`.
validate_args: Python `bool`, default `False`. When `True` distribution
parameters are checked for validity despite possibly degrading runtime
performance. When `False` invalid inputs may silently render incorrect
outputs.
allow_nan_stats: Python `bool`, default `True`. When `True`, statistics
(e.g., mean, mode, variance) use the value "`NaN`" to indicate the
result is undefined. When `False`, an exception is raised if one or
more of the statistic's batch members are undefined.
name: Python `str` name prefixed to Ops created by this class.
"""
with ops.name_scope(name, values=[concentration1, concentration0]) as name:
concentration1 = ops.convert_to_tensor(
concentration1, name="concentration1")
concentration0 = ops.convert_to_tensor(
concentration0, name="concentration0")
super(Kumaraswamy, self).__init__(
distribution=uniform.Uniform(
low=array_ops.zeros([], dtype=concentration1.dtype),
high=array_ops.ones([], dtype=concentration1.dtype),
allow_nan_stats=allow_nan_stats),
bijector=bijectors.Kumaraswamy(
concentration1=concentration1, concentration0=concentration0,
validate_args=validate_args),
batch_shape=distribution_util.get_broadcast_shape(
concentration1, concentration0),
name=name)
self._reparameterization_type = distribution.FULLY_REPARAMETERIZED
def test_all_static_shapes_work(self):
x = array_ops.ones((2, 1, 3))
y = array_ops.ones((1, 5, 3))
z = array_ops.ones(())
self.assertAllEqual([2, 5, 3],
distribution_util.get_broadcast_shape(x, y, z))
def __init__(self,
loc,
scale,
skewness=None,
tailweight=None,
distribution=None,
validate_args=False,
allow_nan_stats=True,
name="SinhArcsinh"):
"""Construct SinhArcsinh distribution on `(-inf, inf)`.
Arguments `(loc, scale, skewness, tailweight)` must have broadcastable shape
(indexing batch dimensions). They must all have the same `dtype`.
Args:
loc: Floating-point `Tensor`.
scale: `Tensor` of same `dtype` as `loc`.
skewness: Skewness parameter. Default is `0.0` (no skew).
tailweight: Tailweight parameter. Default is `1.0` (unchanged tailweight)
distribution: `tf.Distribution`-like instance. Distribution that is
transformed to produce this distribution.
Default is `tf.distributions.Normal(0., 1.)`.
Must be a scalar-batch, scalar-event distribution. Typically
`distribution.reparameterization_type = FULLY_REPARAMETERIZED` or it is
a function of non-trainable parameters. WARNING: If you backprop through
a `SinhArcsinh` sample and `distribution` is not
`FULLY_REPARAMETERIZED` yet is a function of trainable variables, then
the gradient will be incorrect!
validate_args: Python `bool`, default `False`. When `True` distribution
parameters are checked for validity despite possibly degrading runtime
performance. When `False` invalid inputs may silently render incorrect
outputs.
allow_nan_stats: Python `bool`, default `True`. When `True`,
statistics (e.g., mean, mode, variance) use the value "`NaN`" to
indicate the result is undefined. When `False`, an exception is raised
if one or more of the statistic's batch members are undefined.
name: Python `str` name prefixed to Ops created by this class.
"""
parameters = distribution_util.parent_frame_arguments()
with ops.name_scope(name,
values=[loc, scale, skewness, tailweight]) as name:
loc = ops.convert_to_tensor(loc, name="loc")
dtype = loc.dtype
scale = ops.convert_to_tensor(scale, name="scale", dtype=dtype)
tailweight = 1. if tailweight is None else tailweight
has_default_skewness = skewness is None
skewness = 0. if skewness is None else skewness
tailweight = ops.convert_to_tensor(
tailweight, name="tailweight", dtype=dtype)
skewness = ops.convert_to_tensor(skewness, name="skewness", dtype=dtype)
batch_shape = distribution_util.get_broadcast_shape(
loc, scale, tailweight, skewness)
# Recall, with Z a random variable,
# Y := loc + C * F(Z),
# F(Z) := Sinh( (Arcsinh(Z) + skewness) * tailweight )
# F_0(Z) := Sinh( Arcsinh(Z) * tailweight )
# C := 2 * scale / F_0(2)
if distribution is None:
distribution = normal.Normal(
loc=array_ops.zeros([], dtype=dtype),
scale=array_ops.ones([], dtype=dtype),
allow_nan_stats=allow_nan_stats)
else:
asserts = distribution_util.maybe_check_scalar_distribution(
distribution, dtype, validate_args)
if asserts:
loc = control_flow_ops.with_dependencies(asserts, loc)
# Make the SAS bijector, 'F'.
f = bijectors.SinhArcsinh(
skewness=skewness, tailweight=tailweight)
if has_default_skewness:
f_noskew = f
else:
f_noskew = bijectors.SinhArcsinh(
skewness=skewness.dtype.as_numpy_dtype(0.),
tailweight=tailweight)
# Make the AffineScalar bijector, Z --> loc + scale * Z (2 / F_0(2))
c = 2 * scale / f_noskew.forward(ops.convert_to_tensor(2, dtype=dtype))
affine = bijectors.AffineScalar(
shift=loc,
scale=c,
validate_args=validate_args)
bijector = bijectors.Chain([affine, f])
super(SinhArcsinh, self).__init__(
distribution=distribution,
bijector=bijector,
batch_shape=batch_shape,
validate_args=validate_args,
name=name)
self._parameters = parameters
self._loc = loc
self._scale = scale
self._tailweight = tailweight
self._skewness = skewness
def __init__(self,
loc,
scale,
skewness=None,
tailweight=None,
distribution=None,
validate_args=False,
allow_nan_stats=True,
name="SinhArcsinh"):
"""Construct SinhArcsinh distribution on `(-inf, inf)`.
Arguments `(loc, scale, skewness, tailweight)` must have broadcastable shape
(indexing batch dimensions). They must all have the same `dtype`.
Args:
loc: Floating-point `Tensor`.
scale: `Tensor` of same `dtype` as `loc`.
skewness: Skewness parameter. Default is `0.0` (no skew).
tailweight: Tailweight parameter. Default is `1.0` (unchanged tailweight)
distribution: `tf.Distribution`-like instance. Distribution that is
transformed to produce this distribution.
Default is `tf.distributions.Normal(0., 1.)`.
Must be a scalar-batch, scalar-event distribution. Typically
`distribution.reparameterization_type = FULLY_REPARAMETERIZED` or it is
a function of non-trainable parameters. WARNING: If you backprop through
a `SinhArcsinh` sample and `distribution` is not
`FULLY_REPARAMETERIZED` yet is a function of trainable variables, then
the gradient will be incorrect!
validate_args: Python `bool`, default `False`. When `True` distribution
parameters are checked for validity despite possibly degrading runtime
performance. When `False` invalid inputs may silently render incorrect
outputs.
allow_nan_stats: Python `bool`, default `True`. When `True`,
statistics (e.g., mean, mode, variance) use the value "`NaN`" to
indicate the result is undefined. When `False`, an exception is raised
if one or more of the statistic's batch members are undefined.
name: Python `str` name prefixed to Ops created by this class.
"""
parameters = locals()
with ops.name_scope(name, values=[loc, scale, skewness, tailweight]):
loc = ops.convert_to_tensor(loc, name="loc")
dtype = loc.dtype
scale = ops.convert_to_tensor(scale, name="scale", dtype=dtype)
tailweight = 1. if tailweight is None else tailweight
has_default_skewness = skewness is None
skewness = 0. if skewness is None else skewness
tailweight = ops.convert_to_tensor(
tailweight, name="tailweight", dtype=dtype)
skewness = ops.convert_to_tensor(skewness, name="skewness", dtype=dtype)
batch_shape = distribution_util.get_broadcast_shape(
loc, scale, tailweight, skewness)
# Recall, with Z a random variable,
# Y := loc + C * F(Z),
# F(Z) := Sinh( (Arcsinh(Z) + skewness) * tailweight )
# F_0(Z) := Sinh( Arcsinh(Z) * tailweight )
# C := 2 * scale / F_0(2)
if distribution is None:
distribution = normal.Normal(
loc=array_ops.zeros([], dtype=dtype),
scale=array_ops.ones([], dtype=dtype),
allow_nan_stats=allow_nan_stats)
else:
asserts = distribution_util.maybe_check_scalar_distribution(
distribution, dtype, validate_args)
if asserts:
loc = control_flow_ops.with_dependencies(asserts, loc)
# Make the SAS bijector, 'F'.
f = bijectors.SinhArcsinh(
skewness=skewness, tailweight=tailweight)
if has_default_skewness:
f_noskew = f
else:
f_noskew = bijectors.SinhArcsinh(
skewness=skewness.dtype.as_numpy_dtype(0.),
tailweight=tailweight)
# Make the AffineScalar bijector, Z --> loc + scale * Z (2 / F_0(2))
c = 2 * scale / f_noskew.forward(ops.convert_to_tensor(2, dtype=dtype))
affine = bijectors.AffineScalar(
shift=loc,
scale=c,
validate_args=validate_args)
bijector = bijectors.Chain([affine, f])
super(SinhArcsinh, self).__init__(
distribution=distribution,
bijector=bijector,
batch_shape=batch_shape,
validate_args=validate_args,
name=name)
self._parameters = parameters
self._loc = loc
self._scale = scale
self._tailweight = tailweight
self._skewness = skewness
def test_all_static_shapes_work(self):
x = array_ops.ones((2, 1, 3))
y = array_ops.ones((1, 5, 3))
z = array_ops.ones(())
self.assertAllEqual([2, 5, 3],
distribution_util.get_broadcast_shape(x, y, z))
