def testNonInjectiveTransformedDistribution(self):
mu = 1.
sigma = 2.0
abs_normal = self._cls()(distribution=tfd.Normal(loc=mu, scale=sigma),
bijector=tfb.AbsoluteValue(),
validate_args=True)
sp_normal = stats.norm(mu, sigma)
# sample
sample = abs_normal.sample(100000, seed=test_util.test_seed())
self.assertAllEqual([], abs_normal.event_shape)
sample_ = self.evaluate(sample)
self.assertAllEqual([], self.evaluate(abs_normal.event_shape_tensor()))
# Abs > 0, duh!
np.testing.assert_array_less(0, sample_)
# Let X ~ Normal(mu, sigma), Y := |X|, then
# P[Y < 0.77] = P[-0.77 < X < 0.77]
self.assertAllClose(sp_normal.cdf(0.77) - sp_normal.cdf(-0.77),
(sample_ < 0.77).mean(),
rtol=0.01)
# p_Y(y) = p_X(-y) + p_X(y),
self.assertAllClose(
sp_normal.pdf(1.13) + sp_normal.pdf(-1.13),
self.evaluate(abs_normal.prob(1.13)))
# Log[p_Y(y)] = Log[p_X(-y) + p_X(y)]
self.assertAllClose(np.log(sp_normal.pdf(2.13) + sp_normal.pdf(-2.13)),
self.evaluate(abs_normal.log_prob(2.13)))
def testCovarianceNotImplemented(self):
mvn = tfd.MultivariateNormalDiag(loc=[0., 0.], scale_diag=[1., 2.])
# Non-affine bijector.
with self.assertRaisesRegex(
NotImplementedError, '`covariance` is not implemented'):
tfd.TransformedDistribution(
distribution=mvn, bijector=tfb.Exp()).covariance()
# Non-injective bijector.
with self.assertRaisesRegex(
NotImplementedError, '`covariance` is not implemented'):
tfd.TransformedDistribution(
distribution=mvn, bijector=tfb.AbsoluteValue()).covariance()
# Non-vector event shape.
with self.assertRaisesRegex(
NotImplementedError, '`covariance` is only implemented'):
tfd.TransformedDistribution(
distribution=mvn,
bijector=tfb.Reshape(event_shape_out=[2, 1],
event_shape_in=[2])).covariance()
# Multipart bijector.
with self.assertRaisesRegex(
NotImplementedError, '`covariance` is only implemented'):
tfd.TransformedDistribution(
distribution=mvn, bijector=tfb.Split(2)).covariance()
def testCompositeTensor(self):
bijector = tfb.AbsoluteValue(validate_args=True)
flat = tf.nest.flatten(bijector, expand_composites=True)
unflat = tf.nest.pack_sequence_as(bijector, flat, expand_composites=True)
x = tf.convert_to_tensor([0., 1., -1.])
self.assertAllClose(
bijector.forward(x),
tf.function(lambda b_: b_.forward(x))(unflat))
def testBijectorVersusNumpyRewriteOfBasicFunctionsEventNdims0(self):
bijector = tfb.AbsoluteValue(validate_args=True)
self.assertStartsWith(bijector.name, "absolute_value")
x = tf.constant([[0., 1., -1], [0., -5., 3.]]) # Shape [2, 3]
y = tf.math.abs(x)
y_ = self.evaluate(y)
self.assertAllClose(y_, self.evaluate(bijector.forward(x)))
self.assertAllClose((-y_, y_), self.evaluate(bijector.inverse(y)))
self.assertAllClose(
(0., 0.),
self.evaluate(bijector.inverse_log_det_jacobian(y, event_ndims=0)))
# Run things twice to make sure there are no issues in caching the tuples
# returned by .inverse*
self.assertAllClose(y_, self.evaluate(bijector.forward(x)))
self.assertAllClose((-y_, y_), self.evaluate(bijector.inverse(y)))
self.assertAllClose(
(0., 0.),
self.evaluate(bijector.inverse_log_det_jacobian(y, event_ndims=0)))
def testNegativeYRaisesForILDJIfValidateArgs(self):
with self.test_session() as sess:
bijector = tfb.AbsoluteValue(validate_args=True)
with self.assertRaisesOpError("y was negative"):
sess.run(bijector.inverse_log_det_jacobian(-1., event_ndims=0))
def testNegativeYRaisesForInverseIfValidateArgs(self):
with self.test_session() as sess:
bijector = tfb.AbsoluteValue(validate_args=True)
with self.assertRaisesOpError("y was negative"):
sess.run(bijector.inverse(-1.))
