def testFoldl_Simple(self):
elems = constant_op.constant([1, 2, 3, 4, 5, 6], name="data")
r = functional_ops.foldl(
lambda a, x: math_ops.multiply(math_ops.add(a, x), 2), elems)
self.assertAllEqual(208, self.evaluate(r))
r = functional_ops.foldl(
lambda a, x: math_ops.multiply(math_ops.add(a, x), 2),
elems,
initializer=10)
self.assertAllEqual(880, self.evaluate(r))
def testFoldl_Simple(self):
elems = constant_op.constant([1, 2, 3, 4, 5, 6], name="data")
r = functional_ops.foldl(
lambda a, x: math_ops.multiply(math_ops.add(a, x), 2),
elems)
self.assertAllEqual(208, self.evaluate(r))
r = functional_ops.foldl(
lambda a, x: math_ops.multiply(math_ops.add(a, x), 2),
elems,
initializer=10)
self.assertAllEqual(880, self.evaluate(r))
def testFoldl_Simple(self):
with self.test_session():
elems = constant_op.constant([1, 2, 3, 4, 5, 6], name="data")
r = functional_ops.foldl(lambda a, x: math_ops.mul(math_ops.add(a, x), 2),
elems)
self.assertAllEqual(208, r.eval())
r = functional_ops.foldl(
lambda a, x: math_ops.mul(math_ops.add(a, x), 2),
elems,
initializer=10)
self.assertAllEqual(880, r.eval())
def testFoldl_Simple(self):
with self.test_session():
elems = constant_op.constant([1, 2, 3, 4, 5, 6], name="data")
r = functional_ops.foldl(
lambda a, x: math_ops.mul(math_ops.add(a, x), 2), elems)
self.assertAllEqual(208, r.eval())
r = functional_ops.foldl(
lambda a, x: math_ops.mul(math_ops.add(a, x), 2),
elems,
initializer=10)
self.assertAllEqual(880, r.eval())
def testFoldl_MultiInputSingleOutput(self):
with self.test_session():
elems = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0])
initializer = np.array(1.0)
r = functional_ops.foldl(lambda a, x: a + x[0] + x[1], (elems, -elems),
initializer)
self.assertAllEqual(1, self.evaluate(r))
def sparse_boolean_mask(sparse_tensor, mask, name="sparse_boolean_mask"):
"""Boolean mask for `SparseTensor`s.
Args:
sparse_tensor: a `SparseTensor`.
mask: a 1D boolean dense`Tensor` whose length is equal to the 0th dimension
of `sparse_tensor`.
name: optional name for this operation.
Returns:
A `SparseTensor` that contains row `k` of `sparse_tensor` iff `mask[k]` is
`True`.
"""
# TODO(jamieas): consider mask dimension > 1 for symmetry with `boolean_mask`.
with ops.op_scope([sparse_tensor, mask], name):
mask = ops.convert_to_tensor(mask)
mask_rows = array_ops.where(mask)
first_indices = array_ops.squeeze(array_ops.slice(sparse_tensor.indices,
[0, 0], [-1, 1]))
# Identify indices corresponding to the rows identified by mask_rows.
sparse_entry_matches = functional_ops.map_fn(
lambda x: math_ops.equal(first_indices, x),
mask_rows,
dtype=dtypes.bool)
# Combine the rows of index_matches to form a mask for the sparse indices
# and values.
to_retain = array_ops.reshape(
functional_ops.foldl(math_ops.logical_or, sparse_entry_matches), [-1])
return sparse_ops.sparse_retain(sparse_tensor, to_retain)
def foldl(fn, labeled_tensor, initial_value, name=None):
"""Left fold on the list of tensors unpacked from labeled_tensor.
See tf.foldl.
Args:
fn: The function to apply to each unpacked LabeledTensor.
It should have type (LabeledTensor, LabeledTensor) -> LabeledTensor.
Its arguments are (accumulated_value, next_value).
labeled_tensor: The input tensor.
initial_value: The initial value of the accumulator.
name: Optional op name.
Returns:
The accumulated value.
"""
with ops.name_scope(name, 'lt_foldl',
[labeled_tensor, initial_value]) as scope:
labeled_tensor = core.convert_to_labeled_tensor(labeled_tensor)
initial_value = core.convert_to_labeled_tensor(initial_value)
@tc.returns(ops.Tensor)
@tc.accepts(ops.Tensor, ops.Tensor)
def tf_fn(accumulator, next_element):
accumulator_lt = core.LabeledTensor(accumulator, initial_value.axes)
next_element_lt = core.LabeledTensor(
next_element, list(labeled_tensor.axes.values())[1:])
return fn(accumulator_lt, next_element_lt).tensor
foldl_op = functional_ops.foldl(
tf_fn, labeled_tensor.tensor, initializer=initial_value.tensor)
foldl_lt = core.LabeledTensor(foldl_op, initial_value.axes)
return core.identity(foldl_lt, name=scope)
def testFoldl_MultiInputSingleOutput(self):
with self.test_session():
elems = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0])
initializer = np.array(1.0)
r = functional_ops.foldl(lambda a, x: a + x[0] + x[1],
(elems, -elems), initializer)
self.assertAllEqual(1, self.evaluate(r))
def sparse_boolean_mask(sparse_tensor, mask, name="sparse_boolean_mask"):
"""Boolean mask for `SparseTensor`s.
Args:
sparse_tensor: a `SparseTensor`.
mask: a 1D boolean dense`Tensor` whose length is equal to the 0th dimension
of `sparse_tensor`.
name: optional name for this operation.
Returns:
A `SparseTensor` that contains row `k` of `sparse_tensor` iff `mask[k]` is
`True`.
"""
# TODO(jamieas): consider mask dimension > 1 for symmetry with `boolean_mask`.
with ops.op_scope([sparse_tensor, mask], name):
mask = ops.convert_to_tensor(mask)
mask_rows = array_ops.where(mask)
first_indices = array_ops.squeeze(
array_ops.slice(sparse_tensor.indices, [0, 0], [-1, 1]))
# Identify indices corresponding to the rows identified by mask_rows.
sparse_entry_matches = functional_ops.map_fn(
lambda x: math_ops.equal(first_indices, x),
mask_rows,
dtype=dtypes.bool)
# Combine the rows of index_matches to form a mask for the sparse indices
# and values.
to_retain = array_ops.reshape(
functional_ops.foldl(math_ops.logical_or, sparse_entry_matches),
[-1])
return sparse_ops.sparse_retain(sparse_tensor, to_retain)
def testFoldl_MultiInputDifferentDimsSingleOutput(self):
elems = np.array([[1.0, 1.0, 1.0], [2.0, 3.0, 4.0]])
other_elems = np.array([-1.0, 1.0])
initializer = np.array([0.0, 0.0, 0.0])
r = functional_ops.foldl(lambda a, x: a + x[0] * x[1],
(elems, other_elems), initializer)
self.assertAllEqual([1.0, 2.0, 3.0], self.evaluate(r))
def foldl(fn, labeled_tensor, initial_value, name=None):
"""Left fold on the list of tensors unpacked from labeled_tensor.
See tf.foldl.
Args:
fn: The function to apply to each unpacked LabeledTensor.
It should have type (LabeledTensor, LabeledTensor) -> LabeledTensor.
Its arguments are (accumulated_value, next_value).
labeled_tensor: The input tensor.
initial_value: The initial value of the accumulator.
name: Optional op name.
Returns:
The accumulated value.
"""
with ops.name_scope(name, 'lt_foldl',
[labeled_tensor, initial_value]) as scope:
labeled_tensor = core.convert_to_labeled_tensor(labeled_tensor)
initial_value = core.convert_to_labeled_tensor(initial_value)
@tc.returns(ops.Tensor)
@tc.accepts(ops.Tensor, ops.Tensor)
def tf_fn(accumulator, next_element):
accumulator_lt = core.LabeledTensor(accumulator, initial_value.axes)
next_element_lt = core.LabeledTensor(
next_element, list(labeled_tensor.axes.values())[1:])
return fn(accumulator_lt, next_element_lt).tensor
foldl_op = functional_ops.foldl(
tf_fn, labeled_tensor.tensor, initializer=initial_value.tensor)
foldl_lt = core.LabeledTensor(foldl_op, initial_value.axes)
return core.identity(foldl_lt, name=scope)
def testFoldl_MultiInputDifferentDimsSingleOutput(self):
elems = np.array([[1.0, 1.0, 1.0], [2.0, 3.0, 4.0]])
other_elems = np.array([-1.0, 1.0])
initializer = np.array([0.0, 0.0, 0.0])
r = functional_ops.foldl(lambda a, x: a + x[0] * x[1],
(elems, other_elems), initializer)
self.assertAllEqual([1.0, 2.0, 3.0], self.evaluate(r))
def testFoldl_SingleInputMultiOutput(self):
elems = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0])
initializer = np.array([1, -1.0])
r = functional_ops.foldl(lambda a, x: a + x, elems, initializer)
r_value = self.evaluate(r)
self.assertAllEqual(22, r_value[0])
self.assertAllEqual(20, r_value[1])
def testFoldl_SingleInputMultiOutput(self):
elems = np.array([1.0, 2.0, 3.0, 4.0, 5.0, 6.0])
initializer = np.array([1, -1.0])
r = functional_ops.foldl(lambda a, x: a + x, elems, initializer)
r_value = self.evaluate(r)
self.assertAllEqual(22, r_value[0])
self.assertAllEqual(20, r_value[1])
def testFoldShape(self):
x = constant_op.constant([[1, 2, 3], [4, 5, 6]])
def fn(_, current_input):
return current_input
initializer = constant_op.constant([0, 0, 0])
y = functional_ops.foldl(fn, x, initializer=initializer)
self.assertAllEqual(y.get_shape(), self.evaluate(y).shape)
def testFoldl_Scoped(self):
with self.cached_session() as sess:
with variable_scope.variable_scope("root") as varscope:
elems = constant_op.constant([1, 2, 3, 4, 5, 6], name="data")
r = functional_ops.foldl(simple_scoped_fn, elems)
# Check that we have the one variable we asked for here.
self.assertEqual(len(variables.trainable_variables()), 1)
self.assertEqual(variables.trainable_variables()[0].name,
"root/body/two:0")
sess.run([variables.global_variables_initializer()])
self.assertAllEqual(208, self.evaluate(r))
# Now let's reuse our single variable.
varscope.reuse_variables()
r = functional_ops.foldl(simple_scoped_fn, elems, initializer=10)
self.assertEqual(len(variables.trainable_variables()), 1)
self.assertAllEqual(880, self.evaluate(r))
def testFoldShape(self):
x = constant_op.constant([[1, 2, 3], [4, 5, 6]])
def fn(_, current_input):
return current_input
initializer = constant_op.constant([0, 0, 0])
y = functional_ops.foldl(fn, x, initializer=initializer)
self.assertAllEqual(y.get_shape(), self.evaluate(y).shape)
def testFoldl_Scoped(self):
with self.cached_session() as sess:
with variable_scope.variable_scope("root") as varscope:
elems = constant_op.constant([1, 2, 3, 4, 5, 6], name="data")
r = functional_ops.foldl(simple_scoped_fn, elems)
# Check that we have the one variable we asked for here.
self.assertEqual(len(variables.trainable_variables()), 1)
self.assertEqual(variables.trainable_variables()[0].name,
"root/body/two:0")
sess.run([variables.global_variables_initializer()])
self.assertAllEqual(208, self.evaluate(r))
# Now let's reuse our single variable.
varscope.reuse_variables()
r = functional_ops.foldl(simple_scoped_fn, elems, initializer=10)
self.assertEqual(len(variables.trainable_variables()), 1)
self.assertAllEqual(880, self.evaluate(r))
def testFold_Grad(self):
with self.cached_session():
elems = constant_op.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], name="data")
v = constant_op.constant(2.0, name="v")
r = functional_ops.foldl(
lambda a, x: math_ops.multiply(a, x), elems, initializer=v)
r = gradients_impl.gradients(r, v)[0]
self.assertAllEqual(720.0, self.evaluate(r))
r = functional_ops.foldr(
lambda a, x: math_ops.multiply(a, x), elems, initializer=v)
r = gradients_impl.gradients(r, v)[0]
self.assertAllEqual(720.0, self.evaluate(r))
def testFold_Grad(self):
with self.cached_session():
elems = constant_op.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], name="data")
v = constant_op.constant(2.0, name="v")
r = functional_ops.foldl(
lambda a, x: math_ops.multiply(a, x), elems, initializer=v)
r = gradients_impl.gradients(r, v)[0]
self.assertAllEqual(720.0, self.evaluate(r))
r = functional_ops.foldr(
lambda a, x: math_ops.multiply(a, x), elems, initializer=v)
r = gradients_impl.gradients(r, v)[0]
self.assertAllEqual(720.0, self.evaluate(r))
def forward(self, y0, save_intermediate=False):
time_grid = ops.convert_to_tensor(self.ts,
preferred_dtype=float_type,
name='t')
y0 = ops.convert_to_tensor(y0, name='y0')
time_delta_grid = time_grid[1:] - time_grid[:-1]
time_grid = time_grid[1:]
time_combined = tf.concat(
[time_grid[:, None], time_delta_grid[:, None]], axis=1)
scan_func = self._make_scan_func(self.f)
if save_intermediate:
y_grid = functional_ops.scan(scan_func, time_combined, y0)
y_s = array_ops.concat([[y0], y_grid], axis=0)
y_t = y_s[-1, :, :, :]
return y_t, y_s
else:
y_t = functional_ops.foldl(scan_func, time_combined, y0)
return y_t, None
def r_inner(a, x):
return functional_ops.foldl(
lambda b, y: b * y * x, inner_elems, initializer=a)
def r_inner(a, x):
return functional_ops.foldl(
lambda b, y: b * y * x, inner_elems, initializer=a)