def _bincount(
arr,
weights=None,
minlength=None,
maxlength=None, # pylint: disable=unused-argument
dtype=tf.int32,
name=None): # pylint: disable=unused-argument
return np.bincount(arr, weights,
minlength).astype(utils.numpy_dtype(dtype))
def _eye(num_rows,
num_columns=None,
batch_shape=None,
dtype=tf.float32,
name=None): # pylint: disable=unused-argument
dt = utils.numpy_dtype(dtype)
x = np.eye(num_rows, num_columns).astype(dt)
if batch_shape is not None:
x *= np.ones(np.concatenate([batch_shape, [1, 1]], axis=0)).astype(dt)
return x
def _convert_to_tensor(value, dtype=None, dtype_hint=None, name=None): # pylint: disable=unused-argument
"""Emulates tf.convert_to_tensor."""
assert not tf.is_tensor(value), value
if isinstance(value, np.ndarray):
if dtype is not None and value.dtype != utils.numpy_dtype(dtype):
raise ValueError(
'Expected dtype {} but got {} with dtype {}.'.format(
utils.numpy_dtype(dtype), value, value.dtype))
return value
if dtype is None and dtype_hint is not None:
dtype_hint = utils.numpy_dtype(dtype_hint)
value = np.array(value)
# Match TF behavior, which won't downcast e.g. float to int.
if np.issubdtype(value.dtype, np.complex):
if not np.issubdtype(dtype_hint, np.complex):
return value
if np.issubdtype(value.dtype, np.float):
if not np.issubdtype(dtype_hint, np.float):
return value
if np.issubdtype(value.dtype, np.integer):
if not np.issubdtype(dtype_hint, np.integer):
return value
return value.astype(dtype_hint)
return np.array(value, dtype=utils.numpy_dtype(dtype or dtype_hint))
def _zeros_like(input, dtype=None, name=None): # pylint: disable=redefined-builtin
s = _shape(input)
if isinstance(s, (np.ndarray, np.generic)):
return np.zeros(s, utils.numpy_dtype(dtype or input.dtype))
return tf.zeros(s, dtype or s.dtype, name)
def _size(input, out_type=tf.int32, name=None): # pylint: disable=redefined-builtin, unused-argument
return np.prod(np.array(input).shape).astype(utils.numpy_dtype(out_type))
linspace = utils.copy_docstring(
tf.linspace,
lambda start, stop, num, name=None: ( # pylint: disable=g-long-lambda
np.linspace(start, stop, num).astype(np.array(start).dtype)))
meshgrid = utils.copy_docstring(tf.meshgrid, np.meshgrid)
norm = utils.copy_docstring(tf.norm, norm)
one_hot = utils.copy_docstring(tf.one_hot, _one_hot)
ones = utils.copy_docstring(
tf.ones,
lambda shape, dtype=tf.float32, name=None: np.ones( # pylint: disable=g-long-lambda
shape, utils.numpy_dtype(dtype)))
ones_like = utils.copy_docstring(tf.ones_like, _ones_like)
pad = utils.copy_docstring(tf.pad, _pad)
range = utils.copy_docstring( # pylint: disable=redefined-builtin
tf.range,
lambda start, limit=None, delta=1, dtype=None, name='range': ( # pylint: disable=g-long-lambda
np.arange(start, limit, delta, utils.numpy_dtype(dtype))))
rank = utils.copy_docstring(
tf.rank, lambda input, name=None: len(np.array(input).shape)) # pylint: disable=redefined-builtin,g-long-lambda
reshape = utils.copy_docstring(
tf.reshape, lambda tensor, shape, name=None: np.reshape(tensor, shape))
def _top_k(input, k=1, sorted=True, name=None): # pylint: disable=unused-argument,redefined-builtin
raise NotImplementedError
# --- Begin Public Functions --------------------------------------------------
abs = utils.copy_docstring( # pylint: disable=redefined-builtin
tf.math.abs,
lambda x, name=None: np.abs(x))
accumulate_n = utils.copy_docstring(
tf.math.accumulate_n,
lambda inputs, shape=None, tensor_dtype=None, name=None: ( # pylint: disable=g-long-lambda
sum(map(np.array, inputs)).astype(utils.numpy_dtype(tensor_dtype))))
acos = utils.copy_docstring(tf.math.acos, lambda x, name=None: np.arccos(x))
acosh = utils.copy_docstring(tf.math.acosh, lambda x, name=None: np.arccosh(x))
add = utils.copy_docstring(tf.math.add, lambda x, y, name=None: np.add(x, y))
add_n = utils.copy_docstring(
tf.math.add_n, lambda inputs, name=None: sum(map(np.array, inputs)))
angle = utils.copy_docstring(tf.math.angle,
lambda input, name=None: np.angle(input))
argmax = utils.copy_docstring(
tf.math.argmax,
def _convert_to_tensor(value, dtype=None, dtype_hint=None, name=None): # pylint: disable=unused-argument
assert not tf.is_tensor(value), value
return np.array(value, dtype=utils.numpy_dtype(dtype or dtype_hint))
def _constant(value, dtype=None, shape=None, name='Const'): # pylint: disable=unused-argument
x = np.array(value,
dtype=None if dtype is None else utils.numpy_dtype(dtype))
if shape is None:
return x
return np.reshape(x, shape)
return sources
broadcast_dynamic_shape = utils.copy_docstring(tf.broadcast_dynamic_shape,
_broadcast_static_shape)
broadcast_static_shape = utils.copy_docstring(tf.broadcast_static_shape,
_broadcast_static_shape)
broadcast_to = utils.copy_docstring(
tf.broadcast_to,
lambda input, shape, name=None: np.broadcast_to(input, shape))
cast = utils.copy_docstring(
tf.cast,
lambda x, dtype, name=None: np.array(x).astype(utils.numpy_dtype(dtype)))
clip_by_value = utils.copy_docstring(
tf.clip_by_value,
lambda t, clip_value_min, clip_value_max, name=None: # pylint: disable=g-long-lambda
np.clip(t, clip_value_min, clip_value_max))
constant = utils.copy_docstring(tf.constant, _constant)
control_dependencies = utils.copy_docstring(tf.control_dependencies,
_control_dependencies)
convert_to_tensor = utils.copy_docstring(tf.convert_to_tensor,
_convert_to_tensor)
custom_gradient = utils.copy_docstring(tf.custom_gradient, lambda f: f)
# --- Begin Public Functions --------------------------------------------------
broadcast_dynamic_shape = utils.copy_docstring(tf.broadcast_dynamic_shape,
_broadcast_static_shape)
broadcast_static_shape = utils.copy_docstring(tf.broadcast_static_shape,
_broadcast_static_shape)
broadcast_to = utils.copy_docstring(
tf.broadcast_to,
lambda input, shape, name=None: np.broadcast_to(input, shape))
cast = utils.copy_docstring(
tf.cast,
lambda x, dtype, name=None: np.array(x, dtype=utils.numpy_dtype(dtype)))
constant = utils.copy_docstring(tf.constant, _constant)
control_dependencies = utils.copy_docstring(tf.control_dependencies,
_control_dependencies)
convert_to_tensor = utils.copy_docstring(
tf.convert_to_tensor,
lambda value, dtype=None, dtype_hint=None, name=None: ( # pylint: disable=g-long-lambda
np.array(value, dtype=utils.numpy_dtype(dtype or dtype_hint))))
executing_eagerly = utils.copy_docstring(tf.executing_eagerly, lambda: True)
get_static_value = utils.copy_docstring(tf.get_static_value,
lambda tensor, partial=False: tensor)
