def testInvalidTolRaises(self):
loc = rng.rand(2, 3, 4).astype(np.float32)
deterministic = deterministic_lib.VectorDeterministic(
loc, atol=-1, validate_args=True)
with self.cached_session():
with self.assertRaisesOpError("Condition x >= 0"):
deterministic.prob(loc).eval()
def testInvalidTolRaises(self):
loc = rng.rand(2, 3, 4).astype(np.float32)
with self.assertRaisesRegexp(errors.InvalidArgumentError,
"Condition x >= 0"):
_ = deterministic_lib.VectorDeterministic(loc,
atol=-1,
validate_args=True)
def testProbVectorDeterministicWithNoBatchDimsOnRZeroRaisesIfXNotInSameRk(
self):
# 0 batch of deterministics on R^0.
deterministic = deterministic_lib.VectorDeterministic(
[], validate_args=True)
with self.cached_session():
with self.assertRaisesOpError("not defined in the same space"):
deterministic.prob([1.]).eval()
def testProbVectorDeterministicWithNoBatchDims(self):
# 0 batch of deterministics on R^1.
deterministic = deterministic_lib.VectorDeterministic([0.])
with self.cached_session():
self.assertAllClose(1., deterministic.prob([0.]).eval())
self.assertAllClose(0., deterministic.prob([2.]).eval())
self.assertAllClose([1., 0.],
deterministic.prob([[0.], [2.]]).eval())
def testInvalidXRaises(self):
loc = rng.rand(2, 3, 4).astype(np.float32)
deterministic = deterministic_lib.VectorDeterministic(
loc, atol=-1, validate_args=True)
with self.cached_session():
with self.assertRaisesRegexp(ValueError,
"must have rank at least 1"):
deterministic.prob(0.).eval()
def testSampleWithBatchDims(self):
deterministic = deterministic_lib.VectorDeterministic([[0.], [0.]])
for sample_shape in [(), (4, )]:
with self.cached_session():
sample = deterministic.sample(sample_shape)
self.assertAllEqual(sample_shape + (2, 1), sample.get_shape())
self.assertAllClose(
np.zeros(sample_shape + (2, 1)).astype(np.float32),
sample.eval())
def testProbWithNonzeroRTol(self):
# 3 batch of deterministics on R^2.
loc = [[0., 1.], [1., 1.], [100., 100.]]
x = [[0., 1.], [0.9, 1.], [99.9, 100.1]]
deterministic = deterministic_lib.VectorDeterministic(loc, rtol=0.01)
expected_prob = [1., 0., 1.]
with self.cached_session():
prob = deterministic.prob(x)
self.assertAllEqual((3, ), prob.get_shape())
self.assertAllEqual(expected_prob, prob.eval())
def testProbWithNonzeroATol(self):
# 3 batch of deterministics on R^2.
loc = [[0., 1.], [2., 3.], [4., 5.]]
x = [[0., 1.], [1.9, 3.], [3.99, 5.]]
deterministic = deterministic_lib.VectorDeterministic(loc, atol=0.05)
expected_prob = [1., 0., 1.]
with self.cached_session():
prob = deterministic.prob(x)
self.assertAllEqual((3, ), prob.get_shape())
self.assertAllEqual(expected_prob, prob.eval())
def testShape(self):
with self.cached_session():
loc = rng.rand(2, 3, 4)
deterministic = deterministic_lib.VectorDeterministic(loc)
self.assertAllEqual(deterministic.batch_shape_tensor().eval(),
(2, 3))
self.assertAllEqual(deterministic.batch_shape, (2, 3))
self.assertAllEqual(deterministic.event_shape_tensor().eval(), [4])
self.assertEqual(deterministic.event_shape,
tensor_shape.TensorShape([4]))
def testSampleDynamicWithBatchDims(self):
loc = array_ops.placeholder(np.float32)
sample_shape = array_ops.placeholder(np.int32)
deterministic = deterministic_lib.VectorDeterministic(loc)
for sample_shape_ in [(), (4, )]:
with self.cached_session():
sample_ = deterministic.sample(sample_shape).eval(
feed_dict={
loc: [[0.], [0.]],
sample_shape: sample_shape_
})
self.assertAllClose(
np.zeros(sample_shape_ + (2, 1)).astype(np.float32),
sample_)
def testEntropy(self):
loc = np.array([[8.3, 1.2, 3.3], [-0.1, -3.2, 7.]])
deterministic = deterministic_lib.VectorDeterministic(loc=loc)
with self.cached_session() as sess:
entropy_ = sess.run(deterministic.entropy())
self.assertAllEqual(np.zeros(2), entropy_)
def testProbVectorDeterministicWithNoBatchDimsOnRZero(self):
# 0 batch of deterministics on R^0.
deterministic = deterministic_lib.VectorDeterministic(
[], validate_args=True)
with self.cached_session():
self.assertAllClose(1., deterministic.prob([]).eval())
