"""Python wrappers around TensorFlow ops.
This file is MACHINE GENERATED! Do not edit.
Original C++ source file: array_ops.cc
"""
import collections
from tensorflow.python import pywrap_tfe as pywrap_tfe
from tensorflow.python.eager import context as _context
from tensorflow.python.eager import core as _core
from tensorflow.python.eager import execute as _execute
from tensorflow.python.framework import dtypes as _dtypes
from tensorflow.python.framework import op_def_registry as _op_def_registry
from tensorflow.python.framework import ops as _ops
from tensorflow.python.framework import op_def_library as _op_def_library
from tensorflow.python.util.deprecation import deprecated_endpoints
from tensorflow.python.util import dispatch as _dispatch
from tensorflow.python.util.tf_export import tf_export
def batch_matrix_band_part(input, num_lower, num_upper, name=None):
r"""TODO: add doc.
Args:
input: A `Tensor`.
num_lower: A `Tensor` of type `int64`.
num_upper: A `Tensor` of type `int64`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BatchMatrixBandPart", name,
tld.op_callbacks, input, num_lower, num_upper)
return _result
except _core._FallbackException:
try:
return batch_matrix_band_part_eager_fallback(
input, num_lower, num_upper, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BatchMatrixBandPart", input=input, num_lower=num_lower,
num_upper=num_upper, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BatchMatrixBandPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BatchMatrixBandPart = tf_export("raw_ops.BatchMatrixBandPart")(_ops.to_raw_op(batch_matrix_band_part))
def batch_matrix_band_part_eager_fallback(input, num_lower, num_upper, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
num_lower = _ops.convert_to_tensor(num_lower, _dtypes.int64)
num_upper = _ops.convert_to_tensor(num_upper, _dtypes.int64)
_inputs_flat = [input, num_lower, num_upper]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"BatchMatrixBandPart", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BatchMatrixBandPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def batch_matrix_diag(diagonal, name=None):
r"""TODO: add doc.
Args:
diagonal: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `diagonal`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BatchMatrixDiag", name,
tld.op_callbacks, diagonal)
return _result
except _core._FallbackException:
try:
return batch_matrix_diag_eager_fallback(
diagonal, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BatchMatrixDiag", diagonal=diagonal, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BatchMatrixDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BatchMatrixDiag = tf_export("raw_ops.BatchMatrixDiag")(_ops.to_raw_op(batch_matrix_diag))
def batch_matrix_diag_eager_fallback(diagonal, name, ctx):
_attr_T, (diagonal,) = _execute.args_to_matching_eager([diagonal], ctx)
_inputs_flat = [diagonal]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"BatchMatrixDiag", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BatchMatrixDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def batch_matrix_diag_part(input, name=None):
r"""TODO: add doc.
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BatchMatrixDiagPart", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return batch_matrix_diag_part_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BatchMatrixDiagPart", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BatchMatrixDiagPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BatchMatrixDiagPart = tf_export("raw_ops.BatchMatrixDiagPart")(_ops.to_raw_op(batch_matrix_diag_part))
def batch_matrix_diag_part_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"BatchMatrixDiagPart", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BatchMatrixDiagPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def batch_matrix_set_diag(input, diagonal, name=None):
r"""TODO: add doc.
Args:
input: A `Tensor`.
diagonal: A `Tensor`. Must have the same type as `input`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BatchMatrixSetDiag", name,
tld.op_callbacks, input, diagonal)
return _result
except _core._FallbackException:
try:
return batch_matrix_set_diag_eager_fallback(
input, diagonal, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BatchMatrixSetDiag", input=input, diagonal=diagonal, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BatchMatrixSetDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BatchMatrixSetDiag = tf_export("raw_ops.BatchMatrixSetDiag")(_ops.to_raw_op(batch_matrix_set_diag))
def batch_matrix_set_diag_eager_fallback(input, diagonal, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, diagonal], ctx)
(input, diagonal) = _inputs_T
_inputs_flat = [input, diagonal]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"BatchMatrixSetDiag", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BatchMatrixSetDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def batch_to_space(input, crops, block_size, name=None):
r"""BatchToSpace for 4-D tensors of type T.
This is a legacy version of the more general BatchToSpaceND.
Rearranges (permutes) data from batch into blocks of spatial data, followed by
cropping. This is the reverse transformation of SpaceToBatch. More specifically,
this op outputs a copy of the input tensor where values from the `batch`
dimension are moved in spatial blocks to the `height` and `width` dimensions,
followed by cropping along the `height` and `width` dimensions.
Args:
input: A `Tensor`. 4-D tensor with shape
`[batch*block_size*block_size, height_pad/block_size, width_pad/block_size,
depth]`. Note that the batch size of the input tensor must be divisible by
`block_size * block_size`.
crops: A `Tensor`. Must be one of the following types: `int32`, `int64`.
2-D tensor of non-negative integers with shape `[2, 2]`. It specifies
how many elements to crop from the intermediate result across the spatial
dimensions as follows:
crops = [[crop_top, crop_bottom], [crop_left, crop_right]]
block_size: An `int` that is `>= 2`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BatchToSpace", name,
tld.op_callbacks, input, crops, "block_size", block_size)
return _result
except _core._FallbackException:
try:
return batch_to_space_eager_fallback(
input, crops, block_size=block_size, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
block_size = _execute.make_int(block_size, "block_size")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BatchToSpace", input=input, crops=crops, block_size=block_size,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "block_size",
_op._get_attr_int("block_size"), "Tidx",
_op._get_attr_type("Tidx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BatchToSpace", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BatchToSpace = tf_export("raw_ops.BatchToSpace")(_ops.to_raw_op(batch_to_space))
def batch_to_space_eager_fallback(input, crops, block_size, name, ctx):
block_size = _execute.make_int(block_size, "block_size")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tidx, (crops,) = _execute.args_to_matching_eager([crops], ctx, _dtypes.int32)
_inputs_flat = [input, crops]
_attrs = ("T", _attr_T, "block_size", block_size, "Tidx", _attr_Tidx)
_result = _execute.execute(b"BatchToSpace", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BatchToSpace", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export(v1=['batch_to_space_nd', 'manip.batch_to_space_nd'])
@deprecated_endpoints('batch_to_space_nd', 'manip.batch_to_space_nd')
def batch_to_space_nd(input, block_shape, crops, name=None):
r"""BatchToSpace for N-D tensors of type T.
This operation reshapes the "batch" dimension 0 into `M + 1` dimensions of shape
`block_shape + [batch]`, interleaves these blocks back into the grid defined by
the spatial dimensions `[1, ..., M]`, to obtain a result with the same rank as
the input. The spatial dimensions of this intermediate result are then
optionally cropped according to `crops` to produce the output. This is the
reverse of SpaceToBatch. See below for a precise description.
Args:
input: A `Tensor`.
N-D with shape `input_shape = [batch] + spatial_shape + remaining_shape`,
where spatial_shape has M dimensions.
block_shape: A `Tensor`. Must be one of the following types: `int32`, `int64`.
1-D with shape `[M]`, all values must be >= 1.
crops: A `Tensor`. Must be one of the following types: `int32`, `int64`.
2-D with shape `[M, 2]`, all values must be >= 0.
`crops[i] = [crop_start, crop_end]` specifies the amount to crop from input
dimension `i + 1`, which corresponds to spatial dimension `i`. It is
required that
`crop_start[i] + crop_end[i] <= block_shape[i] * input_shape[i + 1]`.
This operation is equivalent to the following steps:
1. Reshape `input` to `reshaped` of shape:
[block_shape[0], ..., block_shape[M-1],
batch / prod(block_shape),
input_shape[1], ..., input_shape[N-1]]
2. Permute dimensions of `reshaped` to produce `permuted` of shape
[batch / prod(block_shape),
input_shape[1], block_shape[0],
...,
input_shape[M], block_shape[M-1],
input_shape[M+1], ..., input_shape[N-1]]
3. Reshape `permuted` to produce `reshaped_permuted` of shape
[batch / prod(block_shape),
input_shape[1] * block_shape[0],
...,
input_shape[M] * block_shape[M-1],
input_shape[M+1],
...,
input_shape[N-1]]
4. Crop the start and end of dimensions `[1, ..., M]` of
`reshaped_permuted` according to `crops` to produce the output of shape:
[batch / prod(block_shape),
input_shape[1] * block_shape[0] - crops[0,0] - crops[0,1],
...,
input_shape[M] * block_shape[M-1] - crops[M-1,0] - crops[M-1,1],
input_shape[M+1], ..., input_shape[N-1]]
Some examples:
(1) For the following input of shape `[4, 1, 1, 1]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [0, 0]]`:
```
[[[[1]]], [[[2]]], [[[3]]], [[[4]]]]
```
The output tensor has shape `[1, 2, 2, 1]` and value:
```
x = [[[[1], [2]], [[3], [4]]]]
```
(2) For the following input of shape `[4, 1, 1, 3]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [0, 0]]`:
```
[[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]]
```
The output tensor has shape `[1, 2, 2, 3]` and value:
```
x = [[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
```
(3) For the following input of shape `[4, 2, 2, 1]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [0, 0]]`:
```
x = [[[[1], [3]], [[9], [11]]],
[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]
```
The output tensor has shape `[1, 4, 4, 1]` and value:
```
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]],
[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
```
(4) For the following input of shape `[8, 1, 3, 1]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [2, 0]]`:
```
x = [[[[0], [1], [3]]], [[[0], [9], [11]]],
[[[0], [2], [4]]], [[[0], [10], [12]]],
[[[0], [5], [7]]], [[[0], [13], [15]]],
[[[0], [6], [8]]], [[[0], [14], [16]]]]
```
The output tensor has shape `[2, 2, 4, 1]` and value:
```
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]]],
[[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
```
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BatchToSpaceND", name,
tld.op_callbacks, input, block_shape, crops)
return _result
except _core._FallbackException:
try:
return batch_to_space_nd_eager_fallback(
input, block_shape, crops, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
batch_to_space_nd, input=input, block_shape=block_shape,
crops=crops, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BatchToSpaceND", input=input, block_shape=block_shape, crops=crops,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
batch_to_space_nd, input=input, block_shape=block_shape,
crops=crops, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tblock_shape",
_op._get_attr_type("Tblock_shape"), "Tcrops",
_op._get_attr_type("Tcrops"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BatchToSpaceND", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BatchToSpaceND = tf_export("raw_ops.BatchToSpaceND")(_ops.to_raw_op(batch_to_space_nd))
def batch_to_space_nd_eager_fallback(input, block_shape, crops, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tblock_shape, (block_shape,) = _execute.args_to_matching_eager([block_shape], ctx, _dtypes.int32)
_attr_Tcrops, (crops,) = _execute.args_to_matching_eager([crops], ctx, _dtypes.int32)
_inputs_flat = [input, block_shape, crops]
_attrs = ("T", _attr_T, "Tblock_shape", _attr_Tblock_shape, "Tcrops",
_attr_Tcrops)
_result = _execute.execute(b"BatchToSpaceND", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BatchToSpaceND", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('bitcast')
def bitcast(input, type, name=None):
r"""Bitcasts a tensor from one type to another without copying data.
Given a tensor `input`, this operation returns a tensor that has the same buffer
data as `input` with datatype `type`.
If the input datatype `T` is larger than the output datatype `type` then the
shape changes from [...] to [..., sizeof(`T`)/sizeof(`type`)].
If `T` is smaller than `type`, the operator requires that the rightmost
dimension be equal to sizeof(`type`)/sizeof(`T`). The shape then goes from
[..., sizeof(`type`)/sizeof(`T`)] to [...].
tf.bitcast() and tf.cast() work differently when real dtype is casted as a complex dtype
(e.g. tf.complex64 or tf.complex128) as tf.cast() make imaginary part 0 while tf.bitcast()
gives module error.
For example,
Example 1:
>>> a = [1., 2., 3.]
>>> equality_bitcast = tf.bitcast(a, tf.complex128)
Traceback (most recent call last):
...
InvalidArgumentError: Cannot bitcast from 1 to 18 [Op:Bitcast]
>>> equality_cast = tf.cast(a, tf.complex128)
>>> print(equality_cast)
tf.Tensor([1.+0.j 2.+0.j 3.+0.j], shape=(3,), dtype=complex128)
Example 2:
>>> tf.bitcast(tf.constant(0xffffffff, dtype=tf.uint32), tf.uint8)
<tf.Tensor: shape=(4,), dtype=uint8, numpy=array([255, 255, 255, 255], dtype=uint8)>
Example 3:
>>> x = [1., 2., 3.]
>>> y = [0., 2., 3.]
>>> equality= tf.equal(x,y)
>>> equality_cast = tf.cast(equality,tf.float32)
>>> equality_bitcast = tf.bitcast(equality_cast,tf.uint8)
>>> print(equality)
tf.Tensor([False True True], shape=(3,), dtype=bool)
>>> print(equality_cast)
tf.Tensor([0. 1. 1.], shape=(3,), dtype=float32)
>>> print(equality_bitcast)
tf.Tensor(
[[ 0 0 0 0]
[ 0 0 128 63]
[ 0 0 128 63]], shape=(3, 4), dtype=uint8)
*NOTE*: Bitcast is implemented as a low-level cast, so machines with different
endian orderings will give different results.
Args:
input: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`, `int64`, `int32`, `uint8`, `uint16`, `uint32`, `uint64`, `int8`, `int16`, `complex64`, `complex128`, `qint8`, `quint8`, `qint16`, `quint16`, `qint32`.
type: A `tf.DType` from: `tf.bfloat16, tf.half, tf.float32, tf.float64, tf.int64, tf.int32, tf.uint8, tf.uint16, tf.uint32, tf.uint64, tf.int8, tf.int16, tf.complex64, tf.complex128, tf.qint8, tf.quint8, tf.qint16, tf.quint16, tf.qint32`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `type`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Bitcast", name,
tld.op_callbacks, input, "type", type)
return _result
except _core._FallbackException:
try:
return bitcast_eager_fallback(
input, type=type, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
bitcast, input=input, type=type, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
type = _execute.make_type(type, "type")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Bitcast", input=input, type=type, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
bitcast, input=input, type=type, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "type",
_op._get_attr_type("type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Bitcast", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Bitcast = tf_export("raw_ops.Bitcast")(_ops.to_raw_op(bitcast))
def bitcast_eager_fallback(input, type, name, ctx):
type = _execute.make_type(type, "type")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "type", type)
_result = _execute.execute(b"Bitcast", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Bitcast", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def broadcast_args(s0, s1, name=None):
r"""Return the shape of s0 op s1 with broadcast.
Given `s0` and `s1`, tensors that represent shapes, compute `r0`, the
broadcasted shape. `s0`, `s1` and `r0` are all integer vectors.
Args:
s0: A `Tensor`. Must be one of the following types: `int32`, `int64`.
s1: A `Tensor`. Must have the same type as `s0`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `s0`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BroadcastArgs", name,
tld.op_callbacks, s0, s1)
return _result
except _core._FallbackException:
try:
return broadcast_args_eager_fallback(
s0, s1, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BroadcastArgs", s0=s0, s1=s1, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BroadcastArgs", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BroadcastArgs = tf_export("raw_ops.BroadcastArgs")(_ops.to_raw_op(broadcast_args))
def broadcast_args_eager_fallback(s0, s1, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([s0, s1], ctx, _dtypes.int32)
(s0, s1) = _inputs_T
_inputs_flat = [s0, s1]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"BroadcastArgs", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BroadcastArgs", _inputs_flat, _attrs, _result)
_result, = _result
return _result
_BroadcastGradientArgsOutput = collections.namedtuple(
"BroadcastGradientArgs",
["r0", "r1"])
def broadcast_gradient_args(s0, s1, name=None):
r"""Return the reduction indices for computing gradients of s0 op s1 with broadcast.
This is typically used by gradient computations for a broadcasting operation.
Args:
s0: A `Tensor`. Must be one of the following types: `int32`, `int64`.
s1: A `Tensor`. Must have the same type as `s0`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (r0, r1).
r0: A `Tensor`. Has the same type as `s0`.
r1: A `Tensor`. Has the same type as `s0`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BroadcastGradientArgs", name,
tld.op_callbacks, s0, s1)
_result = _BroadcastGradientArgsOutput._make(_result)
return _result
except _core._FallbackException:
try:
return broadcast_gradient_args_eager_fallback(
s0, s1, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BroadcastGradientArgs", s0=s0, s1=s1, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BroadcastGradientArgs", _inputs_flat, _attrs, _result)
_result = _BroadcastGradientArgsOutput._make(_result)
return _result
BroadcastGradientArgs = tf_export("raw_ops.BroadcastGradientArgs")(_ops.to_raw_op(broadcast_gradient_args))
def broadcast_gradient_args_eager_fallback(s0, s1, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([s0, s1], ctx, _dtypes.int32)
(s0, s1) = _inputs_T
_inputs_flat = [s0, s1]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"BroadcastGradientArgs", 2, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BroadcastGradientArgs", _inputs_flat, _attrs, _result)
_result = _BroadcastGradientArgsOutput._make(_result)
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('broadcast_to')
def broadcast_to(input, shape, name=None):
r"""Broadcast an array for a compatible shape.
Broadcasting is the process of making arrays to have compatible shapes
for arithmetic operations. Two shapes are compatible if for each
dimension pair they are either equal or one of them is one. When trying
to broadcast a Tensor to a shape, it starts with the trailing dimensions,
and works its way forward.
For example,
>>> x = tf.constant([1, 2, 3])
>>> y = tf.broadcast_to(x, [3, 3])
>>> print(y)
tf.Tensor(
[[1 2 3]
[1 2 3]
[1 2 3]], shape=(3, 3), dtype=int32)
In the above example, the input Tensor with the shape of `[1, 3]`
is broadcasted to output Tensor with shape of `[3, 3]`.
Args:
input: A `Tensor`. A Tensor to broadcast.
shape: A `Tensor`. Must be one of the following types: `int32`, `int64`.
An 1-D `int` Tensor. The shape of the desired output.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "BroadcastTo", name,
tld.op_callbacks, input, shape)
return _result
except _core._FallbackException:
try:
return broadcast_to_eager_fallback(
input, shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
broadcast_to, input=input, shape=shape, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"BroadcastTo", input=input, shape=shape, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
broadcast_to, input=input, shape=shape, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tidx",
_op._get_attr_type("Tidx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"BroadcastTo", _inputs_flat, _attrs, _result)
_result, = _result
return _result
BroadcastTo = tf_export("raw_ops.BroadcastTo")(_ops.to_raw_op(broadcast_to))
def broadcast_to_eager_fallback(input, shape, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tidx, (shape,) = _execute.args_to_matching_eager([shape], ctx, _dtypes.int32)
_inputs_flat = [input, shape]
_attrs = ("T", _attr_T, "Tidx", _attr_Tidx)
_result = _execute.execute(b"BroadcastTo", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"BroadcastTo", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('debugging.check_numerics', v1=['debugging.check_numerics', 'check_numerics'])
@deprecated_endpoints('check_numerics')
def check_numerics(tensor, message, name=None):
r"""Checks a tensor for NaN and Inf values.
When run, reports an `InvalidArgument` error if `tensor` has any values
that are not a number (NaN) or infinity (Inf). Otherwise, passes `tensor` as-is.
Args:
tensor: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`.
message: A `string`. Prefix of the error message.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "CheckNumerics", name,
tld.op_callbacks, tensor, "message", message)
return _result
except _core._FallbackException:
try:
return check_numerics_eager_fallback(
tensor, message=message, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
check_numerics, tensor=tensor, message=message, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
message = _execute.make_str(message, "message")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"CheckNumerics", tensor=tensor, message=message, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
check_numerics, tensor=tensor, message=message, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "message",
_op.get_attr("message"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"CheckNumerics", _inputs_flat, _attrs, _result)
_result, = _result
return _result
CheckNumerics = tf_export("raw_ops.CheckNumerics")(_ops.to_raw_op(check_numerics))
def check_numerics_eager_fallback(tensor, message, name, ctx):
message = _execute.make_str(message, "message")
_attr_T, (tensor,) = _execute.args_to_matching_eager([tensor], ctx)
_inputs_flat = [tensor]
_attrs = ("T", _attr_T, "message", message)
_result = _execute.execute(b"CheckNumerics", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"CheckNumerics", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def check_numerics_v2(tensor, message, name=None):
r"""Checks a tensor for NaN, -Inf and +Inf values.
When run, reports an `InvalidArgument` error if `tensor` has any values
that are not a number (NaN) or infinity (Inf). Otherwise, passes `tensor` as-is.
Unlike CheckNumerics (V1), CheckNumericsV2 distinguishes -Inf and +Inf in the
errors it throws.
Args:
tensor: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`.
message: A `string`. Prefix of the error message.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "CheckNumericsV2", name,
tld.op_callbacks, tensor, "message", message)
return _result
except _core._FallbackException:
try:
return check_numerics_v2_eager_fallback(
tensor, message=message, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
message = _execute.make_str(message, "message")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"CheckNumericsV2", tensor=tensor, message=message, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "message",
_op.get_attr("message"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"CheckNumericsV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
CheckNumericsV2 = tf_export("raw_ops.CheckNumericsV2")(_ops.to_raw_op(check_numerics_v2))
def check_numerics_v2_eager_fallback(tensor, message, name, ctx):
message = _execute.make_str(message, "message")
_attr_T, (tensor,) = _execute.args_to_matching_eager([tensor], ctx)
_inputs_flat = [tensor]
_attrs = ("T", _attr_T, "message", message)
_result = _execute.execute(b"CheckNumericsV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"CheckNumericsV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def concat(concat_dim, values, name=None):
r"""Concatenates tensors along one dimension.
Args:
concat_dim: A `Tensor` of type `int32`.
0-D. The dimension along which to concatenate. Must be in the
range [0, rank(values)).
values: A list of at least 2 `Tensor` objects with the same type.
The `N` Tensors to concatenate. Their ranks and types must match,
and their sizes must match in all dimensions except `concat_dim`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `values`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Concat", name,
tld.op_callbacks, concat_dim, values)
return _result
except _core._FallbackException:
try:
return concat_eager_fallback(
concat_dim, values, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'concat' Op, not %r." % values)
_attr_N = len(values)
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Concat", concat_dim=concat_dim, values=values, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"), "T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Concat", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Concat = tf_export("raw_ops.Concat")(_ops.to_raw_op(concat))
def concat_eager_fallback(concat_dim, values, name, ctx):
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'concat' Op, not %r." % values)
_attr_N = len(values)
_attr_T, values = _execute.args_to_matching_eager(list(values), ctx)
concat_dim = _ops.convert_to_tensor(concat_dim, _dtypes.int32)
_inputs_flat = [concat_dim] + list(values)
_attrs = ("N", _attr_N, "T", _attr_T)
_result = _execute.execute(b"Concat", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Concat", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def concat_offset(concat_dim, shape, name=None):
r"""Computes offsets of concat inputs within its output.
For example:
```
# 'x' is [2, 2, 7]
# 'y' is [2, 3, 7]
# 'z' is [2, 5, 7]
concat_offset(2, [x, y, z]) => [0, 0, 0], [0, 2, 0], [0, 5, 0]
```
This is typically used by gradient computations for a concat operation.
Args:
concat_dim: A `Tensor` of type `int32`.
The dimension along which to concatenate.
shape: A list of at least 2 `Tensor` objects with type `int32`.
The `N` int32 vectors representing shape of tensors being concatenated.
name: A name for the operation (optional).
Returns:
A list with the same length as `shape` of `Tensor` objects with type `int32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ConcatOffset", name,
tld.op_callbacks, concat_dim, shape)
return _result
except _core._FallbackException:
try:
return concat_offset_eager_fallback(
concat_dim, shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(shape, (list, tuple)):
raise TypeError(
"Expected list for 'shape' argument to "
"'concat_offset' Op, not %r." % shape)
_attr_N = len(shape)
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ConcatOffset", concat_dim=concat_dim, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ConcatOffset", _inputs_flat, _attrs, _result)
return _result
ConcatOffset = tf_export("raw_ops.ConcatOffset")(_ops.to_raw_op(concat_offset))
def concat_offset_eager_fallback(concat_dim, shape, name, ctx):
if not isinstance(shape, (list, tuple)):
raise TypeError(
"Expected list for 'shape' argument to "
"'concat_offset' Op, not %r." % shape)
_attr_N = len(shape)
concat_dim = _ops.convert_to_tensor(concat_dim, _dtypes.int32)
shape = _ops.convert_n_to_tensor(shape, _dtypes.int32)
_inputs_flat = [concat_dim] + list(shape)
_attrs = ("N", _attr_N)
_result = _execute.execute(b"ConcatOffset", _attr_N, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ConcatOffset", _inputs_flat, _attrs, _result)
return _result
def concat_v2(values, axis, name=None):
r"""Concatenates tensors along one dimension.
Args:
values: A list of at least 2 `Tensor` objects with the same type.
List of `N` Tensors to concatenate. Their ranks and types must match,
and their sizes must match in all dimensions except `concat_dim`.
axis: A `Tensor`. Must be one of the following types: `int32`, `int64`.
0-D. The dimension along which to concatenate. Must be in the
range [-rank(values), rank(values)).
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `values`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ConcatV2", name,
tld.op_callbacks, values, axis)
return _result
except _core._FallbackException:
try:
return concat_v2_eager_fallback(
values, axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'concat_v2' Op, not %r." % values)
_attr_N = len(values)
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ConcatV2", values=values, axis=axis, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"), "T", _op._get_attr_type("T"),
"Tidx", _op._get_attr_type("Tidx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ConcatV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ConcatV2 = tf_export("raw_ops.ConcatV2")(_ops.to_raw_op(concat_v2))
def concat_v2_eager_fallback(values, axis, name, ctx):
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'concat_v2' Op, not %r." % values)
_attr_N = len(values)
_attr_T, values = _execute.args_to_matching_eager(list(values), ctx)
_attr_Tidx, (axis,) = _execute.args_to_matching_eager([axis], ctx, _dtypes.int32)
_inputs_flat = list(values) + [axis]
_attrs = ("N", _attr_N, "T", _attr_T, "Tidx", _attr_Tidx)
_result = _execute.execute(b"ConcatV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ConcatV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def conjugate_transpose(x, perm, name=None):
r"""Shuffle dimensions of x according to a permutation and conjugate the result.
The output `y` has the same rank as `x`. The shapes of `x` and `y` satisfy:
`y.shape[i] == x.shape[perm[i]] for i in [0, 1, ..., rank(x) - 1]`
`y[i,j,k,...,s,t,u] == conj(x[perm[i], perm[j], perm[k],...,perm[s], perm[t], perm[u]])`
Args:
x: A `Tensor`.
perm: A `Tensor`. Must be one of the following types: `int32`, `int64`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ConjugateTranspose", name,
tld.op_callbacks, x, perm)
return _result
except _core._FallbackException:
try:
return conjugate_transpose_eager_fallback(
x, perm, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ConjugateTranspose", x=x, perm=perm, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tperm",
_op._get_attr_type("Tperm"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ConjugateTranspose", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ConjugateTranspose = tf_export("raw_ops.ConjugateTranspose")(_ops.to_raw_op(conjugate_transpose))
def conjugate_transpose_eager_fallback(x, perm, name, ctx):
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_attr_Tperm, (perm,) = _execute.args_to_matching_eager([perm], ctx, _dtypes.int32)
_inputs_flat = [x, perm]
_attrs = ("T", _attr_T, "Tperm", _attr_Tperm)
_result = _execute.execute(b"ConjugateTranspose", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ConjugateTranspose", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def const(value, dtype, name=None):
r"""Returns a constant tensor.
Args:
value: A `tf.TensorProto`. Attr `value` is the tensor to return.
dtype: A `tf.DType`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `dtype`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Const", name,
tld.op_callbacks, "value", value, "dtype", dtype)
return _result
except _core._FallbackException:
try:
return const_eager_fallback(
value=value, dtype=dtype, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
value = _execute.make_tensor(value, "value")
dtype = _execute.make_type(dtype, "dtype")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Const", value=value, dtype=dtype, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("value", _op.get_attr("value"), "dtype",
_op._get_attr_type("dtype"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Const", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Const = tf_export("raw_ops.Const")(_ops.to_raw_op(const))
def const_eager_fallback(value, dtype, name, ctx):
value = _execute.make_tensor(value, "value")
dtype = _execute.make_type(dtype, "dtype")
_inputs_flat = []
_attrs = ("value", value, "dtype", dtype)
_result = _execute.execute(b"Const", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Const", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def debug_gradient_identity(input, name=None):
r"""Identity op for gradient debugging.
This op is hidden from public in Python. It is used by TensorFlow Debugger to
register gradient tensors for gradient debugging.
This op operates on non-reference-type tensors.
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "DebugGradientIdentity", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return debug_gradient_identity_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"DebugGradientIdentity", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"DebugGradientIdentity", _inputs_flat, _attrs, _result)
_result, = _result
return _result
DebugGradientIdentity = tf_export("raw_ops.DebugGradientIdentity")(_ops.to_raw_op(debug_gradient_identity))
def debug_gradient_identity_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"DebugGradientIdentity", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"DebugGradientIdentity", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def debug_gradient_ref_identity(input, name=None):
r"""Identity op for gradient debugging.
This op is hidden from public in Python. It is used by TensorFlow Debugger to
register gradient tensors for gradient debugging.
This op operates on reference-type tensors.
Args:
input: A mutable `Tensor`.
name: A name for the operation (optional).
Returns:
A mutable `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
raise RuntimeError("debug_gradient_ref_identity op does not support eager execution. Arg 'output' is a ref.")
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"DebugGradientRefIdentity", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"DebugGradientRefIdentity", _inputs_flat, _attrs, _result)
_result, = _result
return _result
DebugGradientRefIdentity = tf_export("raw_ops.DebugGradientRefIdentity")(_ops.to_raw_op(debug_gradient_ref_identity))
def debug_gradient_ref_identity_eager_fallback(input, name, ctx):
raise RuntimeError("debug_gradient_ref_identity op does not support eager execution. Arg 'output' is a ref.")
def deep_copy(x, name=None):
r"""Makes a copy of `x`.
Args:
x: A `Tensor`. The source tensor of type `T`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "DeepCopy", name,
tld.op_callbacks, x)
return _result
except _core._FallbackException:
try:
return deep_copy_eager_fallback(
x, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"DeepCopy", x=x, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"DeepCopy", _inputs_flat, _attrs, _result)
_result, = _result
return _result
DeepCopy = tf_export("raw_ops.DeepCopy")(_ops.to_raw_op(deep_copy))
def deep_copy_eager_fallback(x, name, ctx):
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_inputs_flat = [x]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"DeepCopy", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"DeepCopy", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def depth_to_space(input, block_size, data_format="NHWC", name=None):
r"""DepthToSpace for tensors of type T.
Rearranges data from depth into blocks of spatial data.
This is the reverse transformation of SpaceToDepth. More specifically,
this op outputs a copy of the input tensor where values from the `depth`
dimension are moved in spatial blocks to the `height` and `width` dimensions.
The attr `block_size` indicates the input block size and how the data is moved.
* Chunks of data of size `block_size * block_size` from depth are rearranged
into non-overlapping blocks of size `block_size x block_size`
* The width the output tensor is `input_depth * block_size`, whereas the
height is `input_height * block_size`.
* The Y, X coordinates within each block of the output image are determined
by the high order component of the input channel index.
* The depth of the input tensor must be divisible by
`block_size * block_size`.
The `data_format` attr specifies the layout of the input and output tensors
with the following options:
"NHWC": `[ batch, height, width, channels ]`
"NCHW": `[ batch, channels, height, width ]`
"NCHW_VECT_C":
`qint8 [ batch, channels / 4, height, width, 4 ]`
It is useful to consider the operation as transforming a 6-D Tensor.
e.g. for data_format = NHWC,
Each element in the input tensor can be specified via 6 coordinates,
ordered by decreasing memory layout significance as:
n,iY,iX,bY,bX,oC (where n=batch index, iX, iY means X or Y coordinates
within the input image, bX, bY means coordinates
within the output block, oC means output channels).
The output would be the input transposed to the following layout:
n,iY,bY,iX,bX,oC
This operation is useful for resizing the activations between convolutions
(but keeping all data), e.g. instead of pooling. It is also useful for training
purely convolutional models.
For example, given an input of shape `[1, 1, 1, 4]`, data_format = "NHWC" and
block_size = 2:
```
x = [[[[1, 2, 3, 4]]]]
```
This operation will output a tensor of shape `[1, 2, 2, 1]`:
```
[[[[1], [2]],
[[3], [4]]]]
```
Here, the input has a batch of 1 and each batch element has shape `[1, 1, 4]`,
the corresponding output will have 2x2 elements and will have a depth of
1 channel (1 = `4 / (block_size * block_size)`).
The output element shape is `[2, 2, 1]`.
For an input tensor with larger depth, here of shape `[1, 1, 1, 12]`, e.g.
```
x = [[[[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]]]]
```
This operation, for block size of 2, will return the following tensor of shape
`[1, 2, 2, 3]`
```
[[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
```
Similarly, for the following input of shape `[1 2 2 4]`, and a block size of 2:
```
x = [[[[1, 2, 3, 4],
[5, 6, 7, 8]],
[[9, 10, 11, 12],
[13, 14, 15, 16]]]]
```
the operator will return the following tensor of shape `[1 4 4 1]`:
```
x = [[[ [1], [2], [5], [6]],
[ [3], [4], [7], [8]],
[ [9], [10], [13], [14]],
[ [11], [12], [15], [16]]]]
```
Args:
input: A `Tensor`.
block_size: An `int` that is `>= 2`.
The size of the spatial block, same as in Space2Depth.
data_format: An optional `string` from: `"NHWC", "NCHW", "NCHW_VECT_C"`. Defaults to `"NHWC"`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "DepthToSpace", name,
tld.op_callbacks, input, "block_size", block_size, "data_format",
data_format)
return _result
except _core._FallbackException:
try:
return depth_to_space_eager_fallback(
input, block_size=block_size, data_format=data_format, name=name,
ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
block_size = _execute.make_int(block_size, "block_size")
if data_format is None:
data_format = "NHWC"
data_format = _execute.make_str(data_format, "data_format")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"DepthToSpace", input=input, block_size=block_size,
data_format=data_format, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "block_size",
_op._get_attr_int("block_size"), "data_format",
_op.get_attr("data_format"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"DepthToSpace", _inputs_flat, _attrs, _result)
_result, = _result
return _result
DepthToSpace = tf_export("raw_ops.DepthToSpace")(_ops.to_raw_op(depth_to_space))
def depth_to_space_eager_fallback(input, block_size, data_format, name, ctx):
block_size = _execute.make_int(block_size, "block_size")
if data_format is None:
data_format = "NHWC"
data_format = _execute.make_str(data_format, "data_format")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "block_size", block_size, "data_format",
data_format)
_result = _execute.execute(b"DepthToSpace", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"DepthToSpace", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def dequantize(input, min_range, max_range, mode="MIN_COMBINED", narrow_range=False, axis=-1, dtype=_dtypes.float32, name=None):
r"""Dequantize the 'input' tensor into a float or bfloat16 Tensor.
[min_range, max_range] are scalar floats that specify the range for
the output. The 'mode' attribute controls exactly which calculations are
used to convert the float values to their quantized equivalents.
In 'MIN_COMBINED' mode, each value of the tensor will undergo the following:
```
if T == qint8: in[i] += (range(T) + 1)/ 2.0
out[i] = min_range + (in[i]* (max_range - min_range) / range(T))
```
here `range(T) = numeric_limits<T>::max() - numeric_limits<T>::min()`
*MIN_COMBINED Mode Example*
If the input comes from a QuantizedRelu6, the output type is
quint8 (range of 0-255) but the possible range of QuantizedRelu6 is
0-6. The min_range and max_range values are therefore 0.0 and 6.0.
Dequantize on quint8 will take each value, cast to float, and multiply
by 6 / 255.
Note that if quantizedtype is qint8, the operation will additionally add
each value by 128 prior to casting.
If the mode is 'MIN_FIRST', then this approach is used:
```c++
num_discrete_values = 1 << (# of bits in T)
range_adjust = num_discrete_values / (num_discrete_values - 1)
range = (range_max - range_min) * range_adjust
range_scale = range / num_discrete_values
const double offset_input = static_cast<double>(input) - lowest_quantized;
result = range_min + ((input - numeric_limits<T>::min()) * range_scale)
```
If the mode is `SCALED`, dequantization is performed by multiplying each
input value by a scaling_factor. (Thus an input of 0 always maps to 0.0).
The scaling_factor is determined from `min_range`, `max_range`, and
`narrow_range` in a way that is compatible with `QuantizeAndDequantize{V2|V3}`
and `QuantizeV2`, using the following algorithm:
```c++
const int min_expected_T = std::numeric_limits<T>::min() +
(narrow_range ? 1 : 0);
const int max_expected_T = std::numeric_limits<T>::max();
const float max_expected_T = std::numeric_limits<float>::max();
const float scale_factor =
(std::numeric_limits<T>::min() == 0) ? (max_range / max_expected_T)
: std::max(min_range / min_expected_T,
max_range / max_expected_T);
```
Args:
input: A `Tensor`. Must be one of the following types: `qint8`, `quint8`, `qint32`, `qint16`, `quint16`.
min_range: A `Tensor` of type `float32`.
The minimum scalar value possibly produced for the input.
max_range: A `Tensor` of type `float32`.
The maximum scalar value possibly produced for the input.
mode: An optional `string` from: `"MIN_COMBINED", "MIN_FIRST", "SCALED"`. Defaults to `"MIN_COMBINED"`.
narrow_range: An optional `bool`. Defaults to `False`.
axis: An optional `int`. Defaults to `-1`.
dtype: An optional `tf.DType` from: `tf.bfloat16, tf.float32`. Defaults to `tf.float32`.
Type of the output tensor. Currently Dequantize supports float and bfloat16.
If 'dtype' is 'bfloat16', it only supports 'MIN_COMBINED' mode.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `dtype`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Dequantize", name,
tld.op_callbacks, input, min_range, max_range, "mode", mode,
"narrow_range", narrow_range, "axis", axis, "dtype", dtype)
return _result
except _core._FallbackException:
try:
return dequantize_eager_fallback(
input, min_range, max_range, mode=mode, narrow_range=narrow_range,
axis=axis, dtype=dtype, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if mode is None:
mode = "MIN_COMBINED"
mode = _execute.make_str(mode, "mode")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
if dtype is None:
dtype = _dtypes.float32
dtype = _execute.make_type(dtype, "dtype")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Dequantize", input=input, min_range=min_range, max_range=max_range,
mode=mode, narrow_range=narrow_range, axis=axis,
dtype=dtype, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "mode", _op.get_attr("mode"),
"narrow_range", _op._get_attr_bool("narrow_range"), "axis",
_op._get_attr_int("axis"), "dtype", _op._get_attr_type("dtype"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Dequantize", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Dequantize = tf_export("raw_ops.Dequantize")(_ops.to_raw_op(dequantize))
def dequantize_eager_fallback(input, min_range, max_range, mode, narrow_range, axis, dtype, name, ctx):
if mode is None:
mode = "MIN_COMBINED"
mode = _execute.make_str(mode, "mode")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
if dtype is None:
dtype = _dtypes.float32
dtype = _execute.make_type(dtype, "dtype")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
min_range = _ops.convert_to_tensor(min_range, _dtypes.float32)
max_range = _ops.convert_to_tensor(max_range, _dtypes.float32)
_inputs_flat = [input, min_range, max_range]
_attrs = ("T", _attr_T, "mode", mode, "narrow_range", narrow_range, "axis",
axis, "dtype", dtype)
_result = _execute.execute(b"Dequantize", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Dequantize", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('linalg.tensor_diag', v1=['linalg.tensor_diag', 'diag'])
@deprecated_endpoints('diag')
def diag(diagonal, name=None):
r"""Returns a diagonal tensor with a given diagonal values.
Given a `diagonal`, this operation returns a tensor with the `diagonal` and
everything else padded with zeros. The diagonal is computed as follows:
Assume `diagonal` has dimensions [D1,..., Dk], then the output is a tensor of
rank 2k with dimensions [D1,..., Dk, D1,..., Dk] where:
`output[i1,..., ik, i1,..., ik] = diagonal[i1, ..., ik]` and 0 everywhere else.
For example:
```
# 'diagonal' is [1, 2, 3, 4]
tf.diag(diagonal) ==> [[1, 0, 0, 0]
[0, 2, 0, 0]
[0, 0, 3, 0]
[0, 0, 0, 4]]
```
Args:
diagonal: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
Rank k tensor where k is at most 1.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `diagonal`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Diag", name, tld.op_callbacks,
diagonal)
return _result
except _core._FallbackException:
try:
return diag_eager_fallback(
diagonal, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
diag, diagonal=diagonal, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Diag", diagonal=diagonal, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
diag, diagonal=diagonal, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Diag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Diag = tf_export("raw_ops.Diag")(_ops.to_raw_op(diag))
def diag_eager_fallback(diagonal, name, ctx):
_attr_T, (diagonal,) = _execute.args_to_matching_eager([diagonal], ctx)
_inputs_flat = [diagonal]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Diag", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Diag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('linalg.tensor_diag_part', v1=['linalg.tensor_diag_part', 'diag_part'])
@deprecated_endpoints('diag_part')
def diag_part(input, name=None):
r"""Returns the diagonal part of the tensor.
This operation returns a tensor with the `diagonal` part
of the `input`. The `diagonal` part is computed as follows:
Assume `input` has dimensions `[D1,..., Dk, D1,..., Dk]`, then the output is a
tensor of rank `k` with dimensions `[D1,..., Dk]` where:
`diagonal[i1,..., ik] = input[i1, ..., ik, i1,..., ik]`.
For example:
```
# 'input' is [[1, 0, 0, 0]
[0, 2, 0, 0]
[0, 0, 3, 0]
[0, 0, 0, 4]]
tf.diag_part(input) ==> [1, 2, 3, 4]
```
Args:
input: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
Rank k tensor where k is even and not zero.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "DiagPart", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return diag_part_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
diag_part, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"DiagPart", input=input, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
diag_part, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"DiagPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
DiagPart = tf_export("raw_ops.DiagPart")(_ops.to_raw_op(diag_part))
def diag_part_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"DiagPart", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"DiagPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def edit_distance(hypothesis_indices, hypothesis_values, hypothesis_shape, truth_indices, truth_values, truth_shape, normalize=True, name=None):
r"""Computes the (possibly normalized) Levenshtein Edit Distance.
The inputs are variable-length sequences provided by SparseTensors
(hypothesis_indices, hypothesis_values, hypothesis_shape)
and
(truth_indices, truth_values, truth_shape).
The inputs are:
Args:
hypothesis_indices: A `Tensor` of type `int64`.
The indices of the hypothesis list SparseTensor.
This is an N x R int64 matrix.
hypothesis_values: A `Tensor`.
The values of the hypothesis list SparseTensor.
This is an N-length vector.
hypothesis_shape: A `Tensor` of type `int64`.
The shape of the hypothesis list SparseTensor.
This is an R-length vector.
truth_indices: A `Tensor` of type `int64`.
The indices of the truth list SparseTensor.
This is an M x R int64 matrix.
truth_values: A `Tensor`. Must have the same type as `hypothesis_values`.
The values of the truth list SparseTensor.
This is an M-length vector.
truth_shape: A `Tensor` of type `int64`. truth indices, vector.
normalize: An optional `bool`. Defaults to `True`.
boolean (if true, edit distances are normalized by length of truth).
The output is:
name: A name for the operation (optional).
Returns:
A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "EditDistance", name,
tld.op_callbacks, hypothesis_indices, hypothesis_values,
hypothesis_shape, truth_indices, truth_values, truth_shape,
"normalize", normalize)
return _result
except _core._FallbackException:
try:
return edit_distance_eager_fallback(
hypothesis_indices, hypothesis_values, hypothesis_shape,
truth_indices, truth_values, truth_shape, normalize=normalize,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if normalize is None:
normalize = True
normalize = _execute.make_bool(normalize, "normalize")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"EditDistance", hypothesis_indices=hypothesis_indices,
hypothesis_values=hypothesis_values,
hypothesis_shape=hypothesis_shape,
truth_indices=truth_indices,
truth_values=truth_values, truth_shape=truth_shape,
normalize=normalize, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("normalize", _op._get_attr_bool("normalize"), "T",
_op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"EditDistance", _inputs_flat, _attrs, _result)
_result, = _result
return _result
EditDistance = tf_export("raw_ops.EditDistance")(_ops.to_raw_op(edit_distance))
def edit_distance_eager_fallback(hypothesis_indices, hypothesis_values, hypothesis_shape, truth_indices, truth_values, truth_shape, normalize, name, ctx):
if normalize is None:
normalize = True
normalize = _execute.make_bool(normalize, "normalize")
_attr_T, _inputs_T = _execute.args_to_matching_eager([hypothesis_values, truth_values], ctx)
(hypothesis_values, truth_values) = _inputs_T
hypothesis_indices = _ops.convert_to_tensor(hypothesis_indices, _dtypes.int64)
hypothesis_shape = _ops.convert_to_tensor(hypothesis_shape, _dtypes.int64)
truth_indices = _ops.convert_to_tensor(truth_indices, _dtypes.int64)
truth_shape = _ops.convert_to_tensor(truth_shape, _dtypes.int64)
_inputs_flat = [hypothesis_indices, hypothesis_values, hypothesis_shape, truth_indices, truth_values, truth_shape]
_attrs = ("normalize", normalize, "T", _attr_T)
_result = _execute.execute(b"EditDistance", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"EditDistance", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def empty(shape, dtype, init=False, name=None):
r"""Creates a tensor with the given shape.
This operation creates a tensor of `shape` and `dtype`.
Args:
shape: A `Tensor` of type `int32`.
1-D. Represents the shape of the output tensor.
dtype: A `tf.DType`.
init: An optional `bool`. Defaults to `False`.
If True, initialize the returned tensor with the default value of dtype. Otherwise, the implementation is free not to initializethe tensor's content.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `dtype`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Empty", name,
tld.op_callbacks, shape, "dtype", dtype, "init", init)
return _result
except _core._FallbackException:
try:
return empty_eager_fallback(
shape, dtype=dtype, init=init, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
dtype = _execute.make_type(dtype, "dtype")
if init is None:
init = False
init = _execute.make_bool(init, "init")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Empty", shape=shape, dtype=dtype, init=init, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("dtype", _op._get_attr_type("dtype"), "init",
_op._get_attr_bool("init"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Empty", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Empty = tf_export("raw_ops.Empty")(_ops.to_raw_op(empty))
def empty_eager_fallback(shape, dtype, init, name, ctx):
dtype = _execute.make_type(dtype, "dtype")
if init is None:
init = False
init = _execute.make_bool(init, "init")
shape = _ops.convert_to_tensor(shape, _dtypes.int32)
_inputs_flat = [shape]
_attrs = ("dtype", dtype, "init", init)
_result = _execute.execute(b"Empty", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Empty", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def ensure_shape(input, shape, name=None):
r"""Ensures that the tensor's shape matches the expected shape.
Raises an error if the input tensor's shape does not match the specified shape.
Returns the input tensor otherwise.
Args:
input: A `Tensor`. A tensor, whose shape is to be validated.
shape: A `tf.TensorShape` or list of `ints`.
The expected (possibly partially specified) shape of the input tensor.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "EnsureShape", name,
tld.op_callbacks, input, "shape", shape)
return _result
except _core._FallbackException:
try:
return ensure_shape_eager_fallback(
input, shape=shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
shape = _execute.make_shape(shape, "shape")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"EnsureShape", input=input, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("shape", _op.get_attr("shape"), "T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"EnsureShape", _inputs_flat, _attrs, _result)
_result, = _result
return _result
EnsureShape = tf_export("raw_ops.EnsureShape")(_ops.to_raw_op(ensure_shape))
def ensure_shape_eager_fallback(input, shape, name, ctx):
shape = _execute.make_shape(shape, "shape")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("shape", shape, "T", _attr_T)
_result = _execute.execute(b"EnsureShape", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"EnsureShape", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def expand_dims(input, axis, name=None):
r"""Inserts a dimension of 1 into a tensor's shape.
Given a tensor `input`, this operation inserts a dimension of 1 at the
dimension index `axis` of `input`'s shape. The dimension index `axis` starts at
zero; if you specify a negative number for `axis` it is counted backward from
the end.
This operation is useful if you want to add a batch dimension to a single
element. For example, if you have a single image of shape `[height, width,
channels]`, you can make it a batch of 1 image with `expand_dims(image, 0)`,
which will make the shape `[1, height, width, channels]`.
Other examples:
```
# 't' is a tensor of shape [2]
shape(expand_dims(t, 0)) ==> [1, 2]
shape(expand_dims(t, 1)) ==> [2, 1]
shape(expand_dims(t, -1)) ==> [2, 1]
# 't2' is a tensor of shape [2, 3, 5]
shape(expand_dims(t2, 0)) ==> [1, 2, 3, 5]
shape(expand_dims(t2, 2)) ==> [2, 3, 1, 5]
shape(expand_dims(t2, 3)) ==> [2, 3, 5, 1]
```
This operation requires that:
`-1-input.dims() <= dim <= input.dims()`
This operation is related to `squeeze()`, which removes dimensions of
size 1.
Args:
input: A `Tensor`.
axis: A `Tensor`. Must be one of the following types: `int32`, `int64`.
0-D (scalar). Specifies the dimension index at which to
expand the shape of `input`. Must be in the range
`[-rank(input) - 1, rank(input)]`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ExpandDims", name,
tld.op_callbacks, input, axis)
return _result
except _core._FallbackException:
try:
return expand_dims_eager_fallback(
input, axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ExpandDims", input=input, dim=axis, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tdim",
_op._get_attr_type("Tdim"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ExpandDims", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ExpandDims = tf_export("raw_ops.ExpandDims")(_ops.to_raw_op(expand_dims))
def expand_dims_eager_fallback(input, axis, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tdim, (axis,) = _execute.args_to_matching_eager([axis], ctx, _dtypes.int32)
_inputs_flat = [input, axis]
_attrs = ("T", _attr_T, "Tdim", _attr_Tdim)
_result = _execute.execute(b"ExpandDims", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ExpandDims", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def extract_image_patches(images, ksizes, strides, rates, padding, name=None):
r"""Extract `patches` from `images` and put them in the "depth" output dimension.
Args:
images: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`, `int8`, `int16`, `int32`, `int64`, `uint8`, `uint16`, `uint32`, `uint64`, `complex64`, `complex128`, `bool`.
4-D Tensor with shape `[batch, in_rows, in_cols, depth]`.
ksizes: A list of `ints` that has length `>= 4`.
The size of the sliding window for each dimension of `images`.
strides: A list of `ints` that has length `>= 4`.
How far the centers of two consecutive patches are in
the images. Must be: `[1, stride_rows, stride_cols, 1]`.
rates: A list of `ints` that has length `>= 4`.
Must be: `[1, rate_rows, rate_cols, 1]`. This is the
input stride, specifying how far two consecutive patch samples are in the
input. Equivalent to extracting patches with
`patch_sizes_eff = patch_sizes + (patch_sizes - 1) * (rates - 1)`, followed by
subsampling them spatially by a factor of `rates`. This is equivalent to
`rate` in dilated (a.k.a. Atrous) convolutions.
padding: A `string` from: `"SAME", "VALID"`.
The type of padding algorithm to use.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `images`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ExtractImagePatches", name,
tld.op_callbacks, images, "ksizes", ksizes, "strides", strides,
"rates", rates, "padding", padding)
return _result
except _core._FallbackException:
try:
return extract_image_patches_eager_fallback(
images, ksizes=ksizes, strides=strides, rates=rates,
padding=padding, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(ksizes, (list, tuple)):
raise TypeError(
"Expected list for 'ksizes' argument to "
"'extract_image_patches' Op, not %r." % ksizes)
ksizes = [_execute.make_int(_i, "ksizes") for _i in ksizes]
if not isinstance(strides, (list, tuple)):
raise TypeError(
"Expected list for 'strides' argument to "
"'extract_image_patches' Op, not %r." % strides)
strides = [_execute.make_int(_i, "strides") for _i in strides]
if not isinstance(rates, (list, tuple)):
raise TypeError(
"Expected list for 'rates' argument to "
"'extract_image_patches' Op, not %r." % rates)
rates = [_execute.make_int(_i, "rates") for _i in rates]
padding = _execute.make_str(padding, "padding")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ExtractImagePatches", images=images, ksizes=ksizes, strides=strides,
rates=rates, padding=padding, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("ksizes", _op.get_attr("ksizes"), "strides",
_op.get_attr("strides"), "rates", _op.get_attr("rates"), "T",
_op._get_attr_type("T"), "padding", _op.get_attr("padding"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ExtractImagePatches", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ExtractImagePatches = tf_export("raw_ops.ExtractImagePatches")(_ops.to_raw_op(extract_image_patches))
def extract_image_patches_eager_fallback(images, ksizes, strides, rates, padding, name, ctx):
if not isinstance(ksizes, (list, tuple)):
raise TypeError(
"Expected list for 'ksizes' argument to "
"'extract_image_patches' Op, not %r." % ksizes)
ksizes = [_execute.make_int(_i, "ksizes") for _i in ksizes]
if not isinstance(strides, (list, tuple)):
raise TypeError(
"Expected list for 'strides' argument to "
"'extract_image_patches' Op, not %r." % strides)
strides = [_execute.make_int(_i, "strides") for _i in strides]
if not isinstance(rates, (list, tuple)):
raise TypeError(
"Expected list for 'rates' argument to "
"'extract_image_patches' Op, not %r." % rates)
rates = [_execute.make_int(_i, "rates") for _i in rates]
padding = _execute.make_str(padding, "padding")
_attr_T, (images,) = _execute.args_to_matching_eager([images], ctx)
_inputs_flat = [images]
_attrs = ("ksizes", ksizes, "strides", strides, "rates", rates, "T",
_attr_T, "padding", padding)
_result = _execute.execute(b"ExtractImagePatches", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ExtractImagePatches", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ExtractVolumePatches = tf_export("raw_ops.ExtractVolumePatches")(_ops.to_raw_op(extract_volume_patches))
def extract_volume_patches_eager_fallback(input, ksizes, strides, padding, name, ctx):
if not isinstance(ksizes, (list, tuple)):
raise TypeError(
"Expected list for 'ksizes' argument to "
"'extract_volume_patches' Op, not %r." % ksizes)
ksizes = [_execute.make_int(_i, "ksizes") for _i in ksizes]
if not isinstance(strides, (list, tuple)):
raise TypeError(
"Expected list for 'strides' argument to "
"'extract_volume_patches' Op, not %r." % strides)
strides = [_execute.make_int(_i, "strides") for _i in strides]
padding = _execute.make_str(padding, "padding")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("ksizes", ksizes, "strides", strides, "T", _attr_T, "padding",
padding)
_result = _execute.execute(b"ExtractVolumePatches", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ExtractVolumePatches", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('quantization.fake_quant_with_min_max_args', v1=['quantization.fake_quant_with_min_max_args', 'fake_quant_with_min_max_args'])
@deprecated_endpoints('fake_quant_with_min_max_args')
def fake_quant_with_min_max_args(inputs, min=-6, max=6, num_bits=8, narrow_range=False, name=None):
r"""Fake-quantize the 'inputs' tensor, type float to 'outputs' tensor of same type.
Attributes `[min; max]` define the clamping range for the `inputs` data.
`inputs` values are quantized into the quantization range (`[0; 2^num_bits - 1]`
when `narrow_range` is false and `[1; 2^num_bits - 1]` when it is true) and
then de-quantized and output as floats in `[min; max]` interval.
`num_bits` is the bitwidth of the quantization; between 2 and 16, inclusive.
Before quantization, `min` and `max` values are adjusted with the following
logic.
It is suggested to have `min <= 0 <= max`. If `0` is not in the range of values,
the behavior can be unexpected:
If `0 < min < max`: `min_adj = 0` and `max_adj = max - min`.
If `min < max < 0`: `min_adj = min - max` and `max_adj = 0`.
If `min <= 0 <= max`: `scale = (max - min) / (2^num_bits - 1) `,
`min_adj = scale * round(min / scale)` and `max_adj = max + min_adj - min`.
Quantization is called fake since the output is still in floating point.
Args:
inputs: A `Tensor` of type `float32`.
min: An optional `float`. Defaults to `-6`.
max: An optional `float`. Defaults to `6`.
num_bits: An optional `int`. Defaults to `8`.
narrow_range: An optional `bool`. Defaults to `False`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "FakeQuantWithMinMaxArgs",
name, tld.op_callbacks, inputs, "min", min, "max", max, "num_bits",
num_bits, "narrow_range", narrow_range)
return _result
except _core._FallbackException:
try:
return fake_quant_with_min_max_args_eager_fallback(
inputs, min=min, max=max, num_bits=num_bits,
narrow_range=narrow_range, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_args, inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if min is None:
min = -6
min = _execute.make_float(min, "min")
if max is None:
max = 6
max = _execute.make_float(max, "max")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"FakeQuantWithMinMaxArgs", inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_args, inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("min", _op.get_attr("min"), "max", _op.get_attr("max"),
"num_bits", _op._get_attr_int("num_bits"), "narrow_range",
_op._get_attr_bool("narrow_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"FakeQuantWithMinMaxArgs", _inputs_flat, _attrs, _result)
_result, = _result
return _result
FakeQuantWithMinMaxArgs = tf_export("raw_ops.FakeQuantWithMinMaxArgs")(_ops.to_raw_op(fake_quant_with_min_max_args))
def fake_quant_with_min_max_args_eager_fallback(inputs, min, max, num_bits, narrow_range, name, ctx):
if min is None:
min = -6
min = _execute.make_float(min, "min")
if max is None:
max = 6
max = _execute.make_float(max, "max")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
inputs = _ops.convert_to_tensor(inputs, _dtypes.float32)
_inputs_flat = [inputs]
_attrs = ("min", min, "max", max, "num_bits", num_bits, "narrow_range",
narrow_range)
_result = _execute.execute(b"FakeQuantWithMinMaxArgs", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"FakeQuantWithMinMaxArgs", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('quantization.fake_quant_with_min_max_args_gradient', v1=['quantization.fake_quant_with_min_max_args_gradient', 'fake_quant_with_min_max_args_gradient'])
@deprecated_endpoints('fake_quant_with_min_max_args_gradient')
def fake_quant_with_min_max_args_gradient(gradients, inputs, min=-6, max=6, num_bits=8, narrow_range=False, name=None):
r"""Compute gradients for a FakeQuantWithMinMaxArgs operation.
Args:
gradients: A `Tensor` of type `float32`.
Backpropagated gradients above the FakeQuantWithMinMaxArgs operation.
inputs: A `Tensor` of type `float32`.
Values passed as inputs to the FakeQuantWithMinMaxArgs operation.
min: An optional `float`. Defaults to `-6`.
max: An optional `float`. Defaults to `6`.
num_bits: An optional `int`. Defaults to `8`.
narrow_range: An optional `bool`. Defaults to `False`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name,
"FakeQuantWithMinMaxArgsGradient", name, tld.op_callbacks, gradients,
inputs, "min", min, "max", max, "num_bits", num_bits, "narrow_range",
narrow_range)
return _result
except _core._FallbackException:
try:
return fake_quant_with_min_max_args_gradient_eager_fallback(
gradients, inputs, min=min, max=max, num_bits=num_bits,
narrow_range=narrow_range, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_args_gradient, gradients=gradients,
inputs=inputs, min=min,
max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if min is None:
min = -6
min = _execute.make_float(min, "min")
if max is None:
max = 6
max = _execute.make_float(max, "max")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"FakeQuantWithMinMaxArgsGradient", gradients=gradients, inputs=inputs,
min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_args_gradient, gradients=gradients,
inputs=inputs, min=min,
max=max, num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("min", _op.get_attr("min"), "max", _op.get_attr("max"),
"num_bits", _op._get_attr_int("num_bits"), "narrow_range",
_op._get_attr_bool("narrow_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"FakeQuantWithMinMaxArgsGradient", _inputs_flat, _attrs, _result)
_result, = _result
return _result
FakeQuantWithMinMaxArgsGradient = tf_export("raw_ops.FakeQuantWithMinMaxArgsGradient")(_ops.to_raw_op(fake_quant_with_min_max_args_gradient))
def fake_quant_with_min_max_args_gradient_eager_fallback(gradients, inputs, min, max, num_bits, narrow_range, name, ctx):
if min is None:
min = -6
min = _execute.make_float(min, "min")
if max is None:
max = 6
max = _execute.make_float(max, "max")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
gradients = _ops.convert_to_tensor(gradients, _dtypes.float32)
inputs = _ops.convert_to_tensor(inputs, _dtypes.float32)
_inputs_flat = [gradients, inputs]
_attrs = ("min", min, "max", max, "num_bits", num_bits, "narrow_range",
narrow_range)
_result = _execute.execute(b"FakeQuantWithMinMaxArgsGradient", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"FakeQuantWithMinMaxArgsGradient", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('quantization.fake_quant_with_min_max_vars', v1=['quantization.fake_quant_with_min_max_vars', 'fake_quant_with_min_max_vars'])
@deprecated_endpoints('fake_quant_with_min_max_vars')
def fake_quant_with_min_max_vars(inputs, min, max, num_bits=8, narrow_range=False, name=None):
r"""Fake-quantize the 'inputs' tensor of type float via global float scalars `min`
and `max` to 'outputs' tensor of same shape as `inputs`.
`[min; max]` define the clamping range for the `inputs` data.
`inputs` values are quantized into the quantization range (`[0; 2^num_bits - 1]`
when `narrow_range` is false and `[1; 2^num_bits - 1]` when it is true) and
then de-quantized and output as floats in `[min; max]` interval.
`num_bits` is the bitwidth of the quantization; between 2 and 16, inclusive.
Before quantization, `min` and `max` values are adjusted with the following
logic.
It is suggested to have `min <= 0 <= max`. If `0` is not in the range of values,
the behavior can be unexpected:
If `0 < min < max`: `min_adj = 0` and `max_adj = max - min`.
If `min < max < 0`: `min_adj = min - max` and `max_adj = 0`.
If `min <= 0 <= max`: `scale = (max - min) / (2^num_bits - 1) `,
`min_adj = scale * round(min / scale)` and `max_adj = max + min_adj - min`.
This operation has a gradient and thus allows for training `min` and `max`
values.
Args:
inputs: A `Tensor` of type `float32`.
min: A `Tensor` of type `float32`.
max: A `Tensor` of type `float32`.
num_bits: An optional `int`. Defaults to `8`.
narrow_range: An optional `bool`. Defaults to `False`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "FakeQuantWithMinMaxVars",
name, tld.op_callbacks, inputs, min, max, "num_bits", num_bits,
"narrow_range", narrow_range)
return _result
except _core._FallbackException:
try:
return fake_quant_with_min_max_vars_eager_fallback(
inputs, min, max, num_bits=num_bits, narrow_range=narrow_range,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars, inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"FakeQuantWithMinMaxVars", inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars, inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num_bits", _op._get_attr_int("num_bits"), "narrow_range",
_op._get_attr_bool("narrow_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"FakeQuantWithMinMaxVars", _inputs_flat, _attrs, _result)
_result, = _result
return _result
FakeQuantWithMinMaxVars = tf_export("raw_ops.FakeQuantWithMinMaxVars")(_ops.to_raw_op(fake_quant_with_min_max_vars))
def fake_quant_with_min_max_vars_eager_fallback(inputs, min, max, num_bits, narrow_range, name, ctx):
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
inputs = _ops.convert_to_tensor(inputs, _dtypes.float32)
min = _ops.convert_to_tensor(min, _dtypes.float32)
max = _ops.convert_to_tensor(max, _dtypes.float32)
_inputs_flat = [inputs, min, max]
_attrs = ("num_bits", num_bits, "narrow_range", narrow_range)
_result = _execute.execute(b"FakeQuantWithMinMaxVars", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"FakeQuantWithMinMaxVars", _inputs_flat, _attrs, _result)
_result, = _result
return _result
_FakeQuantWithMinMaxVarsGradientOutput = collections.namedtuple(
"FakeQuantWithMinMaxVarsGradient",
["backprops_wrt_input", "backprop_wrt_min", "backprop_wrt_max"])
@_dispatch.add_dispatch_list
@tf_export('quantization.fake_quant_with_min_max_vars_gradient', v1=['quantization.fake_quant_with_min_max_vars_gradient', 'fake_quant_with_min_max_vars_gradient'])
@deprecated_endpoints('fake_quant_with_min_max_vars_gradient')
def fake_quant_with_min_max_vars_gradient(gradients, inputs, min, max, num_bits=8, narrow_range=False, name=None):
r"""Compute gradients for a FakeQuantWithMinMaxVars operation.
Args:
gradients: A `Tensor` of type `float32`.
Backpropagated gradients above the FakeQuantWithMinMaxVars operation.
inputs: A `Tensor` of type `float32`.
Values passed as inputs to the FakeQuantWithMinMaxVars operation.
min, max: Quantization interval, scalar floats.
min: A `Tensor` of type `float32`.
max: A `Tensor` of type `float32`.
num_bits: An optional `int`. Defaults to `8`.
The bitwidth of the quantization; between 2 and 8, inclusive.
narrow_range: An optional `bool`. Defaults to `False`.
Whether to quantize into 2^num_bits - 1 distinct values.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (backprops_wrt_input, backprop_wrt_min, backprop_wrt_max).
backprops_wrt_input: A `Tensor` of type `float32`.
backprop_wrt_min: A `Tensor` of type `float32`.
backprop_wrt_max: A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name,
"FakeQuantWithMinMaxVarsGradient", name, tld.op_callbacks, gradients,
inputs, min, max, "num_bits", num_bits, "narrow_range", narrow_range)
_result = _FakeQuantWithMinMaxVarsGradientOutput._make(_result)
return _result
except _core._FallbackException:
try:
return fake_quant_with_min_max_vars_gradient_eager_fallback(
gradients, inputs, min, max, num_bits=num_bits,
narrow_range=narrow_range, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars_gradient, gradients=gradients,
inputs=inputs, min=min,
max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"FakeQuantWithMinMaxVarsGradient", gradients=gradients, inputs=inputs,
min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars_gradient, gradients=gradients,
inputs=inputs, min=min,
max=max, num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num_bits", _op._get_attr_int("num_bits"), "narrow_range",
_op._get_attr_bool("narrow_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"FakeQuantWithMinMaxVarsGradient", _inputs_flat, _attrs, _result)
_result = _FakeQuantWithMinMaxVarsGradientOutput._make(_result)
return _result
FakeQuantWithMinMaxVarsGradient = tf_export("raw_ops.FakeQuantWithMinMaxVarsGradient")(_ops.to_raw_op(fake_quant_with_min_max_vars_gradient))
def fake_quant_with_min_max_vars_gradient_eager_fallback(gradients, inputs, min, max, num_bits, narrow_range, name, ctx):
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
gradients = _ops.convert_to_tensor(gradients, _dtypes.float32)
inputs = _ops.convert_to_tensor(inputs, _dtypes.float32)
min = _ops.convert_to_tensor(min, _dtypes.float32)
max = _ops.convert_to_tensor(max, _dtypes.float32)
_inputs_flat = [gradients, inputs, min, max]
_attrs = ("num_bits", num_bits, "narrow_range", narrow_range)
_result = _execute.execute(b"FakeQuantWithMinMaxVarsGradient", 3,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"FakeQuantWithMinMaxVarsGradient", _inputs_flat, _attrs, _result)
_result = _FakeQuantWithMinMaxVarsGradientOutput._make(_result)
return _result
@_dispatch.add_dispatch_list
@tf_export('quantization.fake_quant_with_min_max_vars_per_channel', v1=['quantization.fake_quant_with_min_max_vars_per_channel', 'fake_quant_with_min_max_vars_per_channel'])
@deprecated_endpoints('fake_quant_with_min_max_vars_per_channel')
def fake_quant_with_min_max_vars_per_channel(inputs, min, max, num_bits=8, narrow_range=False, name=None):
r"""Fake-quantize the 'inputs' tensor of type float and one of the shapes: `[d]`,
`[b, d]` `[b, h, w, d]` via per-channel floats `min` and `max` of shape `[d]`
to 'outputs' tensor of same shape as `inputs`.
`[min; max]` define the clamping range for the `inputs` data.
`inputs` values are quantized into the quantization range (`[0; 2^num_bits - 1]`
when `narrow_range` is false and `[1; 2^num_bits - 1]` when it is true) and
then de-quantized and output as floats in `[min; max]` interval.
`num_bits` is the bitwidth of the quantization; between 2 and 16, inclusive.
Before quantization, `min` and `max` values are adjusted with the following
logic.
It is suggested to have `min <= 0 <= max`. If `0` is not in the range of values,
the behavior can be unexpected:
If `0 < min < max`: `min_adj = 0` and `max_adj = max - min`.
If `min < max < 0`: `min_adj = min - max` and `max_adj = 0`.
If `min <= 0 <= max`: `scale = (max - min) / (2^num_bits - 1) `,
`min_adj = scale * round(min / scale)` and `max_adj = max + min_adj - min`.
This operation has a gradient and thus allows for training `min` and `max`
values.
Args:
inputs: A `Tensor` of type `float32`.
min: A `Tensor` of type `float32`.
max: A `Tensor` of type `float32`.
num_bits: An optional `int`. Defaults to `8`.
narrow_range: An optional `bool`. Defaults to `False`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name,
"FakeQuantWithMinMaxVarsPerChannel", name, tld.op_callbacks, inputs,
min, max, "num_bits", num_bits, "narrow_range", narrow_range)
return _result
except _core._FallbackException:
try:
return fake_quant_with_min_max_vars_per_channel_eager_fallback(
inputs, min, max, num_bits=num_bits, narrow_range=narrow_range,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars_per_channel, inputs=inputs,
min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"FakeQuantWithMinMaxVarsPerChannel", inputs=inputs, min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars_per_channel, inputs=inputs, min=min,
max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num_bits", _op._get_attr_int("num_bits"), "narrow_range",
_op._get_attr_bool("narrow_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"FakeQuantWithMinMaxVarsPerChannel", _inputs_flat, _attrs, _result)
_result, = _result
return _result
FakeQuantWithMinMaxVarsPerChannel = tf_export("raw_ops.FakeQuantWithMinMaxVarsPerChannel")(_ops.to_raw_op(fake_quant_with_min_max_vars_per_channel))
def fake_quant_with_min_max_vars_per_channel_eager_fallback(inputs, min, max, num_bits, narrow_range, name, ctx):
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
inputs = _ops.convert_to_tensor(inputs, _dtypes.float32)
min = _ops.convert_to_tensor(min, _dtypes.float32)
max = _ops.convert_to_tensor(max, _dtypes.float32)
_inputs_flat = [inputs, min, max]
_attrs = ("num_bits", num_bits, "narrow_range", narrow_range)
_result = _execute.execute(b"FakeQuantWithMinMaxVarsPerChannel", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"FakeQuantWithMinMaxVarsPerChannel", _inputs_flat, _attrs, _result)
_result, = _result
return _result
_FakeQuantWithMinMaxVarsPerChannelGradientOutput = collections.namedtuple(
"FakeQuantWithMinMaxVarsPerChannelGradient",
["backprops_wrt_input", "backprop_wrt_min", "backprop_wrt_max"])
@_dispatch.add_dispatch_list
@tf_export('quantization.fake_quant_with_min_max_vars_per_channel_gradient', v1=['quantization.fake_quant_with_min_max_vars_per_channel_gradient', 'fake_quant_with_min_max_vars_per_channel_gradient'])
@deprecated_endpoints('fake_quant_with_min_max_vars_per_channel_gradient')
def fake_quant_with_min_max_vars_per_channel_gradient(gradients, inputs, min, max, num_bits=8, narrow_range=False, name=None):
r"""Compute gradients for a FakeQuantWithMinMaxVarsPerChannel operation.
Args:
gradients: A `Tensor` of type `float32`.
Backpropagated gradients above the FakeQuantWithMinMaxVars operation,
shape one of: `[d]`, `[b, d]`, `[b, h, w, d]`.
inputs: A `Tensor` of type `float32`.
Values passed as inputs to the FakeQuantWithMinMaxVars operation, shape
same as `gradients`.
min, max: Quantization interval, floats of shape `[d]`.
min: A `Tensor` of type `float32`.
max: A `Tensor` of type `float32`.
num_bits: An optional `int`. Defaults to `8`.
The bitwidth of the quantization; between 2 and 16, inclusive.
narrow_range: An optional `bool`. Defaults to `False`.
Whether to quantize into 2^num_bits - 1 distinct values.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (backprops_wrt_input, backprop_wrt_min, backprop_wrt_max).
backprops_wrt_input: A `Tensor` of type `float32`.
backprop_wrt_min: A `Tensor` of type `float32`.
backprop_wrt_max: A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name,
"FakeQuantWithMinMaxVarsPerChannelGradient", name, tld.op_callbacks,
gradients, inputs, min, max, "num_bits", num_bits, "narrow_range",
narrow_range)
_result = _FakeQuantWithMinMaxVarsPerChannelGradientOutput._make(_result)
return _result
except _core._FallbackException:
try:
return fake_quant_with_min_max_vars_per_channel_gradient_eager_fallback(
gradients, inputs, min, max, num_bits=num_bits,
narrow_range=narrow_range, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars_per_channel_gradient, gradients=gradients,
inputs=inputs,
min=min,
max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"FakeQuantWithMinMaxVarsPerChannelGradient", gradients=gradients,
inputs=inputs, min=min,
max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
fake_quant_with_min_max_vars_per_channel_gradient, gradients=gradients,
inputs=inputs,
min=min, max=max,
num_bits=num_bits,
narrow_range=narrow_range,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num_bits", _op._get_attr_int("num_bits"), "narrow_range",
_op._get_attr_bool("narrow_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"FakeQuantWithMinMaxVarsPerChannelGradient", _inputs_flat, _attrs, _result)
_result = _FakeQuantWithMinMaxVarsPerChannelGradientOutput._make(_result)
return _result
FakeQuantWithMinMaxVarsPerChannelGradient = tf_export("raw_ops.FakeQuantWithMinMaxVarsPerChannelGradient")(_ops.to_raw_op(fake_quant_with_min_max_vars_per_channel_gradient))
def fake_quant_with_min_max_vars_per_channel_gradient_eager_fallback(gradients, inputs, min, max, num_bits, narrow_range, name, ctx):
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
gradients = _ops.convert_to_tensor(gradients, _dtypes.float32)
inputs = _ops.convert_to_tensor(inputs, _dtypes.float32)
min = _ops.convert_to_tensor(min, _dtypes.float32)
max = _ops.convert_to_tensor(max, _dtypes.float32)
_inputs_flat = [gradients, inputs, min, max]
_attrs = ("num_bits", num_bits, "narrow_range", narrow_range)
_result = _execute.execute(b"FakeQuantWithMinMaxVarsPerChannelGradient", 3,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"FakeQuantWithMinMaxVarsPerChannelGradient", _inputs_flat, _attrs, _result)
_result = _FakeQuantWithMinMaxVarsPerChannelGradientOutput._make(_result)
return _result
def fill(dims, value, name=None):
r"""Creates a tensor filled with a scalar value.
This operation creates a tensor of shape `dims` and fills it with `value`.
For example:
```
# Output tensor has shape [2, 3].
fill([2, 3], 9) ==> [[9, 9, 9]
[9, 9, 9]]
```
`tf.fill` differs from `tf.constant` in a few ways:
* `tf.fill` only supports scalar contents, whereas `tf.constant` supports
Tensor values.
* `tf.fill` creates an Op in the computation graph that constructs the actual
Tensor value at runtime. This is in contrast to `tf.constant` which embeds
the entire Tensor into the graph with a `Const` node.
* Because `tf.fill` evaluates at graph runtime, it supports dynamic shapes
based on other runtime Tensors, unlike `tf.constant`.
Args:
dims: A `Tensor`. Must be one of the following types: `int32`, `int64`.
1-D. Represents the shape of the output tensor.
value: A `Tensor`. 0-D (scalar). Value to fill the returned tensor.
@compatibility(numpy)
Equivalent to np.full
@end_compatibility
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `value`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Fill", name, tld.op_callbacks,
dims, value)
return _result
except _core._FallbackException:
try:
return fill_eager_fallback(
dims, value, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Fill", dims=dims, value=value, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "index_type",
_op._get_attr_type("index_type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Fill", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Fill = tf_export("raw_ops.Fill")(_ops.to_raw_op(fill))
def fill_eager_fallback(dims, value, name, ctx):
_attr_T, (value,) = _execute.args_to_matching_eager([value], ctx)
_attr_index_type, (dims,) = _execute.args_to_matching_eager([dims], ctx, _dtypes.int32)
_inputs_flat = [dims, value]
_attrs = ("T", _attr_T, "index_type", _attr_index_type)
_result = _execute.execute(b"Fill", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Fill", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def fingerprint(data, method, name=None):
r"""Generates fingerprint values.
Generates fingerprint values of `data`.
Fingerprint op considers the first dimension of `data` as the batch dimension,
and `output[i]` contains the fingerprint value generated from contents in
`data[i, ...]` for all `i`.
Fingerprint op writes fingerprint values as byte arrays. For example, the
default method `farmhash64` generates a 64-bit fingerprint value at a time.
This 8-byte value is written out as an `uint8` array of size 8, in little-endian
order.
For example, suppose that `data` has data type `DT_INT32` and shape (2, 3, 4),
and that the fingerprint method is `farmhash64`. In this case, the output shape
is (2, 8), where 2 is the batch dimension size of `data`, and 8 is the size of
each fingerprint value in bytes. `output[0, :]` is generated from 12 integers in
`data[0, :, :]` and similarly `output[1, :]` is generated from other 12 integers
in `data[1, :, :]`.
Note that this op fingerprints the raw underlying buffer, and it does not
fingerprint Tensor's metadata such as data type and/or shape. For example, the
fingerprint values are invariant under reshapes and bitcasts as long as the
batch dimension remain the same:
```
Fingerprint(data) == Fingerprint(Reshape(data, ...))
Fingerprint(data) == Fingerprint(Bitcast(data, ...))
```
For string data, one should expect `Fingerprint(data) !=
Fingerprint(ReduceJoin(data))` in general.
Args:
data: A `Tensor`. Must have rank 1 or higher.
method: A `Tensor` of type `string`.
Fingerprint method used by this op. Currently available method is
`farmhash::fingerprint64`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `uint8`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Fingerprint", name,
tld.op_callbacks, data, method)
return _result
except _core._FallbackException:
try:
return fingerprint_eager_fallback(
data, method, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Fingerprint", data=data, method=method, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Fingerprint", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Fingerprint = tf_export("raw_ops.Fingerprint")(_ops.to_raw_op(fingerprint))
def fingerprint_eager_fallback(data, method, name, ctx):
_attr_T, (data,) = _execute.args_to_matching_eager([data], ctx)
method = _ops.convert_to_tensor(method, _dtypes.string)
_inputs_flat = [data, method]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Fingerprint", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Fingerprint", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def gather(params, indices, validate_indices=True, name=None):
r"""Gather slices from `params` according to `indices`.
`indices` must be an integer tensor of any dimension (usually 0-D or 1-D).
Produces an output tensor with shape `indices.shape + params.shape[1:]` where:
```python
# Scalar indices
output[:, ..., :] = params[indices, :, ... :]
# Vector indices
output[i, :, ..., :] = params[indices[i], :, ... :]
# Higher rank indices
output[i, ..., j, :, ... :] = params[indices[i, ..., j], :, ..., :]
```
If `indices` is a permutation and `len(indices) == params.shape[0]` then
this operation will permute `params` accordingly.
`validate_indices`: DEPRECATED. If this operation is assigned to CPU, values in
`indices` are always validated to be within range. If assigned to GPU,
out-of-bound indices result in safe but unspecified behavior, which may include
raising an error.
<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="https://www.tensorflow.org/images/Gather.png" alt>
</div>
Args:
params: A `Tensor`.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
validate_indices: An optional `bool`. Defaults to `True`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `params`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Gather", name,
tld.op_callbacks, params, indices, "validate_indices",
validate_indices)
return _result
except _core._FallbackException:
try:
return gather_eager_fallback(
params, indices, validate_indices=validate_indices, name=name,
ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if validate_indices is None:
validate_indices = True
validate_indices = _execute.make_bool(validate_indices, "validate_indices")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Gather", params=params, indices=indices,
validate_indices=validate_indices, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("validate_indices", _op._get_attr_bool("validate_indices"),
"Tparams", _op._get_attr_type("Tparams"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Gather", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Gather = tf_export("raw_ops.Gather")(_ops.to_raw_op(gather))
def gather_eager_fallback(params, indices, validate_indices, name, ctx):
if validate_indices is None:
validate_indices = True
validate_indices = _execute.make_bool(validate_indices, "validate_indices")
_attr_Tparams, (params,) = _execute.args_to_matching_eager([params], ctx)
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_inputs_flat = [params, indices]
_attrs = ("validate_indices", validate_indices, "Tparams", _attr_Tparams,
"Tindices", _attr_Tindices)
_result = _execute.execute(b"Gather", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Gather", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def gather_nd(params, indices, name=None):
r"""Gather slices from `params` into a Tensor with shape specified by `indices`.
`indices` is a K-dimensional integer tensor, best thought of as a
(K-1)-dimensional tensor of indices into `params`, where each element defines a
slice of `params`:
output[\\(i_0, ..., i_{K-2}\\)] = params[indices[\\(i_0, ..., i_{K-2}\\)]]
Whereas in `tf.gather` `indices` defines slices into the `axis`
dimension of `params`, in `tf.gather_nd`, `indices` defines slices into the
first `N` dimensions of `params`, where `N = indices.shape[-1]`.
The last dimension of `indices` can be at most the rank of
`params`:
indices.shape[-1] <= params.rank
The last dimension of `indices` corresponds to elements
(if `indices.shape[-1] == params.rank`) or slices
(if `indices.shape[-1] < params.rank`) along dimension `indices.shape[-1]`
of `params`. The output tensor has shape
indices.shape[:-1] + params.shape[indices.shape[-1]:]
Note that on CPU, if an out of bound index is found, an error is returned.
On GPU, if an out of bound index is found, a 0 is stored in the
corresponding output value.
Some examples below.
Simple indexing into a matrix:
```python
indices = [[0, 0], [1, 1]]
params = [['a', 'b'], ['c', 'd']]
output = ['a', 'd']
```
Slice indexing into a matrix:
```python
indices = [[1], [0]]
params = [['a', 'b'], ['c', 'd']]
output = [['c', 'd'], ['a', 'b']]
```
Indexing into a 3-tensor:
```python
indices = [[1]]
params = [[['a0', 'b0'], ['c0', 'd0']],
[['a1', 'b1'], ['c1', 'd1']]]
output = [[['a1', 'b1'], ['c1', 'd1']]]
indices = [[0, 1], [1, 0]]
params = [[['a0', 'b0'], ['c0', 'd0']],
[['a1', 'b1'], ['c1', 'd1']]]
output = [['c0', 'd0'], ['a1', 'b1']]
indices = [[0, 0, 1], [1, 0, 1]]
params = [[['a0', 'b0'], ['c0', 'd0']],
[['a1', 'b1'], ['c1', 'd1']]]
output = ['b0', 'b1']
```
Batched indexing into a matrix:
```python
indices = [[[0, 0]], [[0, 1]]]
params = [['a', 'b'], ['c', 'd']]
output = [['a'], ['b']]
```
Batched slice indexing into a matrix:
```python
indices = [[[1]], [[0]]]
params = [['a', 'b'], ['c', 'd']]
output = [[['c', 'd']], [['a', 'b']]]
```
Batched indexing into a 3-tensor:
```python
indices = [[[1]], [[0]]]
params = [[['a0', 'b0'], ['c0', 'd0']],
[['a1', 'b1'], ['c1', 'd1']]]
output = [[[['a1', 'b1'], ['c1', 'd1']]],
[[['a0', 'b0'], ['c0', 'd0']]]]
indices = [[[0, 1], [1, 0]], [[0, 0], [1, 1]]]
params = [[['a0', 'b0'], ['c0', 'd0']],
[['a1', 'b1'], ['c1', 'd1']]]
output = [[['c0', 'd0'], ['a1', 'b1']],
[['a0', 'b0'], ['c1', 'd1']]]
indices = [[[0, 0, 1], [1, 0, 1]], [[0, 1, 1], [1, 1, 0]]]
params = [[['a0', 'b0'], ['c0', 'd0']],
[['a1', 'b1'], ['c1', 'd1']]]
output = [['b0', 'b1'], ['d0', 'c1']]
```
See also `tf.gather` and `tf.batch_gather`.
Args:
params: A `Tensor`. The tensor from which to gather values.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Index tensor.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `params`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "GatherNd", name,
tld.op_callbacks, params, indices)
return _result
except _core._FallbackException:
try:
return gather_nd_eager_fallback(
params, indices, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"GatherNd", params=params, indices=indices, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("Tparams", _op._get_attr_type("Tparams"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"GatherNd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
GatherNd = tf_export("raw_ops.GatherNd")(_ops.to_raw_op(gather_nd))
def gather_nd_eager_fallback(params, indices, name, ctx):
_attr_Tparams, (params,) = _execute.args_to_matching_eager([params], ctx)
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_inputs_flat = [params, indices]
_attrs = ("Tparams", _attr_Tparams, "Tindices", _attr_Tindices)
_result = _execute.execute(b"GatherNd", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"GatherNd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def gather_v2(params, indices, axis, batch_dims=0, name=None):
r"""Gather slices from `params` axis `axis` according to `indices`.
`indices` must be an integer tensor of any dimension (usually 0-D or 1-D).
Produces an output tensor with shape `params.shape[:axis] + indices.shape +
params.shape[axis + 1:]` where:
```python
# Scalar indices (output is rank(params) - 1).
output[a_0, ..., a_n, b_0, ..., b_n] =
params[a_0, ..., a_n, indices, b_0, ..., b_n]
# Vector indices (output is rank(params)).
output[a_0, ..., a_n, i, b_0, ..., b_n] =
params[a_0, ..., a_n, indices[i], b_0, ..., b_n]
# Higher rank indices (output is rank(params) + rank(indices) - 1).
output[a_0, ..., a_n, i, ..., j, b_0, ... b_n] =
params[a_0, ..., a_n, indices[i, ..., j], b_0, ..., b_n]
```
<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="https://www.tensorflow.org/images/Gather.png" alt>
</div>
Note that on CPU, if an out of bound index is found, an error is returned.
On GPU, if an out of bound index is found, a 0 is stored in the
corresponding output value.
See also `tf.batch_gather` and `tf.gather_nd`.
Args:
params: A `Tensor`.
The tensor from which to gather values. Must be at least rank
`axis + 1`.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Index tensor. Must be in range `[0, params.shape[axis])`.
axis: A `Tensor`. Must be one of the following types: `int32`, `int64`.
The axis in `params` to gather `indices` from. Defaults to the first
dimension. Supports negative indexes.
batch_dims: An optional `int`. Defaults to `0`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `params`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "GatherV2", name,
tld.op_callbacks, params, indices, axis, "batch_dims", batch_dims)
return _result
except _core._FallbackException:
try:
return gather_v2_eager_fallback(
params, indices, axis, batch_dims=batch_dims, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if batch_dims is None:
batch_dims = 0
batch_dims = _execute.make_int(batch_dims, "batch_dims")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"GatherV2", params=params, indices=indices, axis=axis,
batch_dims=batch_dims, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("batch_dims", _op._get_attr_int("batch_dims"), "Tparams",
_op._get_attr_type("Tparams"), "Tindices",
_op._get_attr_type("Tindices"), "Taxis",
_op._get_attr_type("Taxis"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"GatherV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
GatherV2 = tf_export("raw_ops.GatherV2")(_ops.to_raw_op(gather_v2))
def gather_v2_eager_fallback(params, indices, axis, batch_dims, name, ctx):
if batch_dims is None:
batch_dims = 0
batch_dims = _execute.make_int(batch_dims, "batch_dims")
_attr_Tparams, (params,) = _execute.args_to_matching_eager([params], ctx)
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_attr_Taxis, (axis,) = _execute.args_to_matching_eager([axis], ctx)
_inputs_flat = [params, indices, axis]
_attrs = ("batch_dims", batch_dims, "Tparams", _attr_Tparams, "Tindices",
_attr_Tindices, "Taxis", _attr_Taxis)
_result = _execute.execute(b"GatherV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"GatherV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('guarantee_const')
def guarantee_const(input, name=None):
r"""Gives a guarantee to the TF runtime that the input tensor is a constant.
The runtime is then free to make optimizations based on this.
Only accepts value typed tensors as inputs and rejects resource variable handles
as input.
Returns the input tensor without modification.
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "GuaranteeConst", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return guarantee_const_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
guarantee_const, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"GuaranteeConst", input=input, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
guarantee_const, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"GuaranteeConst", _inputs_flat, _attrs, _result)
_result, = _result
return _result
GuaranteeConst = tf_export("raw_ops.GuaranteeConst")(_ops.to_raw_op(guarantee_const))
def guarantee_const_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"GuaranteeConst", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"GuaranteeConst", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def identity(input, name=None):
r"""Return a tensor with the same shape and contents as the input tensor or value.
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Identity", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return identity_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Identity", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Identity", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Identity = tf_export("raw_ops.Identity")(_ops.to_raw_op(identity))
def identity_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Identity", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Identity", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('identity_n')
def identity_n(input, name=None):
r"""Returns a list of tensors with the same shapes and contents as the input
tensors.
This op can be used to override the gradient for complicated functions. For
example, suppose y = f(x) and we wish to apply a custom function g for backprop
such that dx = g(dy). In Python,
```python
with tf.get_default_graph().gradient_override_map(
{'IdentityN': 'OverrideGradientWithG'}):
y, _ = identity_n([f(x), x])
@tf.RegisterGradient('OverrideGradientWithG')
def ApplyG(op, dy, _):
return [None, g(dy)] # Do not backprop to f(x).
```
Args:
input: A list of `Tensor` objects.
name: A name for the operation (optional).
Returns:
A list of `Tensor` objects. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "IdentityN", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return identity_n_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
identity_n, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"IdentityN", input=input, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
identity_n, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op.get_attr("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"IdentityN", _inputs_flat, _attrs, _result)
return _result
IdentityN = tf_export("raw_ops.IdentityN")(_ops.to_raw_op(identity_n))
def identity_n_eager_fallback(input, name, ctx):
_attr_T, input = _execute.convert_to_mixed_eager_tensors(input, ctx)
_inputs_flat = list(input)
_attrs = ("T", _attr_T)
_result = _execute.execute(b"IdentityN", len(input), inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"IdentityN", _inputs_flat, _attrs, _result)
return _result
def immutable_const(dtype, shape, memory_region_name, name=None):
r"""Returns immutable tensor from memory region.
The current implementation memmaps the tensor from a file.
Args:
dtype: A `tf.DType`. Type of the returned tensor.
shape: A `tf.TensorShape` or list of `ints`. Shape of the returned tensor.
memory_region_name: A `string`.
Name of readonly memory region used by the tensor, see
NewReadOnlyMemoryRegionFromFile in tensorflow::Env.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `dtype`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ImmutableConst", name,
tld.op_callbacks, "dtype", dtype, "shape", shape,
"memory_region_name", memory_region_name)
return _result
except _core._FallbackException:
try:
return immutable_const_eager_fallback(
dtype=dtype, shape=shape, memory_region_name=memory_region_name,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
dtype = _execute.make_type(dtype, "dtype")
shape = _execute.make_shape(shape, "shape")
memory_region_name = _execute.make_str(memory_region_name, "memory_region_name")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ImmutableConst", dtype=dtype, shape=shape,
memory_region_name=memory_region_name, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("dtype", _op._get_attr_type("dtype"), "shape",
_op.get_attr("shape"), "memory_region_name",
_op.get_attr("memory_region_name"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ImmutableConst", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ImmutableConst = tf_export("raw_ops.ImmutableConst")(_ops.to_raw_op(immutable_const))
def immutable_const_eager_fallback(dtype, shape, memory_region_name, name, ctx):
dtype = _execute.make_type(dtype, "dtype")
shape = _execute.make_shape(shape, "shape")
memory_region_name = _execute.make_str(memory_region_name, "memory_region_name")
_inputs_flat = []
_attrs = ("dtype", dtype, "shape", shape, "memory_region_name",
memory_region_name)
_result = _execute.execute(b"ImmutableConst", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ImmutableConst", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def inplace_add(x, i, v, name=None):
r""" Adds v into specified rows of x.
Computes y = x; y[i, :] += v; return y.
Args:
x: A `Tensor`. A `Tensor` of type T.
i: A `Tensor` of type `int32`.
A vector. Indices into the left-most dimension of `x`.
v: A `Tensor`. Must have the same type as `x`.
A `Tensor` of type T. Same dimension sizes as x except the first dimension, which must be the same as i's size.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "InplaceAdd", name,
tld.op_callbacks, x, i, v)
return _result
except _core._FallbackException:
try:
return inplace_add_eager_fallback(
x, i, v, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"InplaceAdd", x=x, i=i, v=v, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"InplaceAdd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
InplaceAdd = tf_export("raw_ops.InplaceAdd")(_ops.to_raw_op(inplace_add))
def inplace_add_eager_fallback(x, i, v, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([x, v], ctx)
(x, v) = _inputs_T
i = _ops.convert_to_tensor(i, _dtypes.int32)
_inputs_flat = [x, i, v]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"InplaceAdd", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"InplaceAdd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def inplace_sub(x, i, v, name=None):
r""" Subtracts `v` into specified rows of `x`.
Computes y = x; y[i, :] -= v; return y.
Args:
x: A `Tensor`. A `Tensor` of type T.
i: A `Tensor` of type `int32`.
A vector. Indices into the left-most dimension of `x`.
v: A `Tensor`. Must have the same type as `x`.
A `Tensor` of type T. Same dimension sizes as x except the first dimension, which must be the same as i's size.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "InplaceSub", name,
tld.op_callbacks, x, i, v)
return _result
except _core._FallbackException:
try:
return inplace_sub_eager_fallback(
x, i, v, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"InplaceSub", x=x, i=i, v=v, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"InplaceSub", _inputs_flat, _attrs, _result)
_result, = _result
return _result
InplaceSub = tf_export("raw_ops.InplaceSub")(_ops.to_raw_op(inplace_sub))
def inplace_sub_eager_fallback(x, i, v, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([x, v], ctx)
(x, v) = _inputs_T
i = _ops.convert_to_tensor(i, _dtypes.int32)
_inputs_flat = [x, i, v]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"InplaceSub", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"InplaceSub", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def inplace_update(x, i, v, name=None):
r""" Updates specified rows with values in `v`.
Computes `x[i, :] = v; return x`.
Args:
x: A `Tensor`. A tensor of type `T`.
i: A `Tensor` of type `int32`.
A vector. Indices into the left-most dimension of `x`.
v: A `Tensor`. Must have the same type as `x`.
A `Tensor` of type T. Same dimension sizes as x except the first dimension, which must be the same as i's size.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "InplaceUpdate", name,
tld.op_callbacks, x, i, v)
return _result
except _core._FallbackException:
try:
return inplace_update_eager_fallback(
x, i, v, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"InplaceUpdate", x=x, i=i, v=v, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"InplaceUpdate", _inputs_flat, _attrs, _result)
_result, = _result
return _result
InplaceUpdate = tf_export("raw_ops.InplaceUpdate")(_ops.to_raw_op(inplace_update))
def inplace_update_eager_fallback(x, i, v, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([x, v], ctx)
(x, v) = _inputs_T
i = _ops.convert_to_tensor(i, _dtypes.int32)
_inputs_flat = [x, i, v]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"InplaceUpdate", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"InplaceUpdate", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('math.invert_permutation', v1=['math.invert_permutation', 'invert_permutation'])
@deprecated_endpoints('invert_permutation')
def invert_permutation(x, name=None):
r"""Computes the inverse permutation of a tensor.
This operation computes the inverse of an index permutation. It takes a 1-D
integer tensor `x`, which represents the indices of a zero-based array, and
swaps each value with its index position. In other words, for an output tensor
`y` and an input tensor `x`, this operation computes the following:
`y[x[i]] = i for i in [0, 1, ..., len(x) - 1]`
The values must include 0. There can be no duplicate values or negative values.
For example:
```
# tensor `x` is [3, 4, 0, 2, 1]
invert_permutation(x) ==> [2, 4, 3, 0, 1]
```
Args:
x: A `Tensor`. Must be one of the following types: `int32`, `int64`. 1-D.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "InvertPermutation", name,
tld.op_callbacks, x)
return _result
except _core._FallbackException:
try:
return invert_permutation_eager_fallback(
x, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
invert_permutation, x=x, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"InvertPermutation", x=x, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
invert_permutation, x=x, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"InvertPermutation", _inputs_flat, _attrs, _result)
_result, = _result
return _result
InvertPermutation = tf_export("raw_ops.InvertPermutation")(_ops.to_raw_op(invert_permutation))
def invert_permutation_eager_fallback(x, name, ctx):
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx, _dtypes.int32)
_inputs_flat = [x]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"InvertPermutation", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"InvertPermutation", _inputs_flat, _attrs, _result)
_result, = _result
return _result
_ListDiffOutput = collections.namedtuple(
"ListDiff",
["out", "idx"])
def list_diff(x, y, out_idx=_dtypes.int32, name=None):
r"""Computes the difference between two lists of numbers or strings.
Given a list `x` and a list `y`, this operation returns a list `out` that
represents all values that are in `x` but not in `y`. The returned list `out`
is sorted in the same order that the numbers appear in `x` (duplicates are
preserved). This operation also returns a list `idx` that represents the
position of each `out` element in `x`. In other words:
`out[i] = x[idx[i]] for i in [0, 1, ..., len(out) - 1]`
For example, given this input:
```
x = [1, 2, 3, 4, 5, 6]
y = [1, 3, 5]
```
This operation would return:
```
out ==> [2, 4, 6]
idx ==> [1, 3, 5]
```
Args:
x: A `Tensor`. 1-D. Values to keep.
y: A `Tensor`. Must have the same type as `x`. 1-D. Values to remove.
out_idx: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (out, idx).
out: A `Tensor`. Has the same type as `x`.
idx: A `Tensor` of type `out_idx`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ListDiff", name,
tld.op_callbacks, x, y, "out_idx", out_idx)
_result = _ListDiffOutput._make(_result)
return _result
except _core._FallbackException:
try:
return list_diff_eager_fallback(
x, y, out_idx=out_idx, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ListDiff", x=x, y=y, out_idx=out_idx, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_idx",
_op._get_attr_type("out_idx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ListDiff", _inputs_flat, _attrs, _result)
_result = _ListDiffOutput._make(_result)
return _result
ListDiff = tf_export("raw_ops.ListDiff")(_ops.to_raw_op(list_diff))
def list_diff_eager_fallback(x, y, out_idx, name, ctx):
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_attr_T, _inputs_T = _execute.args_to_matching_eager([x, y], ctx)
(x, y) = _inputs_T
_inputs_flat = [x, y]
_attrs = ("T", _attr_T, "out_idx", out_idx)
_result = _execute.execute(b"ListDiff", 2, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ListDiff", _inputs_flat, _attrs, _result)
_result = _ListDiffOutput._make(_result)
return _result
def lower_bound(sorted_inputs, values, out_type=_dtypes.int32, name=None):
r"""Applies lower_bound(sorted_search_values, values) along each row.
Each set of rows with the same index in (sorted_inputs, values) is treated
independently. The resulting row is the equivalent of calling
`np.searchsorted(sorted_inputs, values, side='left')`.
The result is not a global index to the entire
`Tensor`, but rather just the index in the last dimension.
A 2-D example:
sorted_sequence = [[0, 3, 9, 9, 10],
[1, 2, 3, 4, 5]]
values = [[2, 4, 9],
[0, 2, 6]]
result = LowerBound(sorted_sequence, values)
result == [[1, 2, 2],
[0, 1, 5]]
Args:
sorted_inputs: A `Tensor`. 2-D Tensor where each row is ordered.
values: A `Tensor`. Must have the same type as `sorted_inputs`.
2-D Tensor with the same numbers of rows as `sorted_search_values`. Contains
the values that will be searched for in `sorted_search_values`.
out_type: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `out_type`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "LowerBound", name,
tld.op_callbacks, sorted_inputs, values, "out_type", out_type)
return _result
except _core._FallbackException:
try:
return lower_bound_eager_fallback(
sorted_inputs, values, out_type=out_type, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"LowerBound", sorted_inputs=sorted_inputs, values=values,
out_type=out_type, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_type",
_op._get_attr_type("out_type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"LowerBound", _inputs_flat, _attrs, _result)
_result, = _result
return _result
LowerBound = tf_export("raw_ops.LowerBound")(_ops.to_raw_op(lower_bound))
def lower_bound_eager_fallback(sorted_inputs, values, out_type, name, ctx):
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_attr_T, _inputs_T = _execute.args_to_matching_eager([sorted_inputs, values], ctx)
(sorted_inputs, values) = _inputs_T
_inputs_flat = [sorted_inputs, values]
_attrs = ("T", _attr_T, "out_type", out_type)
_result = _execute.execute(b"LowerBound", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"LowerBound", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('linalg.band_part', v1=['linalg.band_part', 'matrix_band_part'])
@deprecated_endpoints('matrix_band_part')
def matrix_band_part(input, num_lower, num_upper, name=None):
r"""Copy a tensor setting everything outside a central band in each innermost matrix
to zero.
The `band` part is computed as follows:
Assume `input` has `k` dimensions `[I, J, K, ..., M, N]`, then the output is a
tensor with the same shape where
`band[i, j, k, ..., m, n] = in_band(m, n) * input[i, j, k, ..., m, n]`.
The indicator function
`in_band(m, n) = (num_lower < 0 || (m-n) <= num_lower)) &&
(num_upper < 0 || (n-m) <= num_upper)`.
For example:
```
# if 'input' is [[ 0, 1, 2, 3]
[-1, 0, 1, 2]
[-2, -1, 0, 1]
[-3, -2, -1, 0]],
tf.matrix_band_part(input, 1, -1) ==> [[ 0, 1, 2, 3]
[-1, 0, 1, 2]
[ 0, -1, 0, 1]
[ 0, 0, -1, 0]],
tf.matrix_band_part(input, 2, 1) ==> [[ 0, 1, 0, 0]
[-1, 0, 1, 0]
[-2, -1, 0, 1]
[ 0, -2, -1, 0]]
```
Useful special cases:
```
tf.matrix_band_part(input, 0, -1) ==> Upper triangular part.
tf.matrix_band_part(input, -1, 0) ==> Lower triangular part.
tf.matrix_band_part(input, 0, 0) ==> Diagonal.
```
Args:
input: A `Tensor`. Rank `k` tensor.
num_lower: A `Tensor`. Must be one of the following types: `int32`, `int64`.
0-D tensor. Number of subdiagonals to keep. If negative, keep entire
lower triangle.
num_upper: A `Tensor`. Must have the same type as `num_lower`.
0-D tensor. Number of superdiagonals to keep. If negative, keep
entire upper triangle.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixBandPart", name,
tld.op_callbacks, input, num_lower, num_upper)
return _result
except _core._FallbackException:
try:
return matrix_band_part_eager_fallback(
input, num_lower, num_upper, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
matrix_band_part, input=input, num_lower=num_lower,
num_upper=num_upper, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixBandPart", input=input, num_lower=num_lower,
num_upper=num_upper, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
matrix_band_part, input=input, num_lower=num_lower,
num_upper=num_upper, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tindex",
_op._get_attr_type("Tindex"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixBandPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixBandPart = tf_export("raw_ops.MatrixBandPart")(_ops.to_raw_op(matrix_band_part))
def matrix_band_part_eager_fallback(input, num_lower, num_upper, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tindex, _inputs_Tindex = _execute.args_to_matching_eager([num_lower, num_upper], ctx, _dtypes.int64)
(num_lower, num_upper) = _inputs_Tindex
_inputs_flat = [input, num_lower, num_upper]
_attrs = ("T", _attr_T, "Tindex", _attr_Tindex)
_result = _execute.execute(b"MatrixBandPart", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixBandPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_diag(diagonal, name=None):
r"""Returns a batched diagonal tensor with a given batched diagonal values.
Given a `diagonal`, this operation returns a tensor with the `diagonal` and
everything else padded with zeros. The diagonal is computed as follows:
Assume `diagonal` has `k` dimensions `[I, J, K, ..., N]`, then the output is a
tensor of rank `k+1` with dimensions [I, J, K, ..., N, N]` where:
`output[i, j, k, ..., m, n] = 1{m=n} * diagonal[i, j, k, ..., n]`.
For example:
```
# 'diagonal' is [[1, 2, 3, 4], [5, 6, 7, 8]]
and diagonal.shape = (2, 4)
tf.matrix_diag(diagonal) ==> [[[1, 0, 0, 0]
[0, 2, 0, 0]
[0, 0, 3, 0]
[0, 0, 0, 4]],
[[5, 0, 0, 0]
[0, 6, 0, 0]
[0, 0, 7, 0]
[0, 0, 0, 8]]]
which has shape (2, 4, 4)
```
Args:
diagonal: A `Tensor`. Rank `k`, where `k >= 1`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `diagonal`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixDiag", name,
tld.op_callbacks, diagonal)
return _result
except _core._FallbackException:
try:
return matrix_diag_eager_fallback(
diagonal, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixDiag", diagonal=diagonal, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixDiag = tf_export("raw_ops.MatrixDiag")(_ops.to_raw_op(matrix_diag))
def matrix_diag_eager_fallback(diagonal, name, ctx):
_attr_T, (diagonal,) = _execute.args_to_matching_eager([diagonal], ctx)
_inputs_flat = [diagonal]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"MatrixDiag", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_diag_part(input, name=None):
r"""Returns the batched diagonal part of a batched tensor.
This operation returns a tensor with the `diagonal` part
of the batched `input`. The `diagonal` part is computed as follows:
Assume `input` has `k` dimensions `[I, J, K, ..., M, N]`, then the output is a
tensor of rank `k - 1` with dimensions `[I, J, K, ..., min(M, N)]` where:
`diagonal[i, j, k, ..., n] = input[i, j, k, ..., n, n]`.
The input must be at least a matrix.
For example:
```
# 'input' is [[[1, 0, 0, 0]
[0, 2, 0, 0]
[0, 0, 3, 0]
[0, 0, 0, 4]],
[[5, 0, 0, 0]
[0, 6, 0, 0]
[0, 0, 7, 0]
[0, 0, 0, 8]]]
and input.shape = (2, 4, 4)
tf.matrix_diag_part(input) ==> [[1, 2, 3, 4], [5, 6, 7, 8]]
which has shape (2, 4)
```
Args:
input: A `Tensor`. Rank `k` tensor where `k >= 2`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixDiagPart", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return matrix_diag_part_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixDiagPart", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixDiagPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixDiagPart = tf_export("raw_ops.MatrixDiagPart")(_ops.to_raw_op(matrix_diag_part))
def matrix_diag_part_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"MatrixDiagPart", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixDiagPart", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_diag_part_v2(input, k, padding_value, name=None):
r"""Returns the batched diagonal part of a batched tensor.
Returns a tensor with the `k[0]`-th to `k[1]`-th diagonals of the batched
`input`.
Assume `input` has `r` dimensions `[I, J, ..., L, M, N]`.
Let `max_diag_len` be the maximum length among all diagonals to be extracted,
`max_diag_len = min(M + min(k[1], 0), N + min(-k[0], 0))`
Let `num_diags` be the number of diagonals to extract,
`num_diags = k[1] - k[0] + 1`.
If `num_diags == 1`, the output tensor is of rank `r - 1` with shape
`[I, J, ..., L, max_diag_len]` and values:
```
diagonal[i, j, ..., l, n]
= input[i, j, ..., l, n+y, n+x] ; if 0 <= n+y < M and 0 <= n+x < N,
padding_value ; otherwise.
```
where `y = max(-k[1], 0)`, `x = max(k[1], 0)`.
Otherwise, the output tensor has rank `r` with dimensions
`[I, J, ..., L, num_diags, max_diag_len]` with values:
```
diagonal[i, j, ..., l, m, n]
= input[i, j, ..., l, n+y, n+x] ; if 0 <= n+y < M and 0 <= n+x < N,
padding_value ; otherwise.
```
where `d = k[1] - m`, `y = max(-d, 0)`, and `x = max(d, 0)`.
The input must be at least a matrix.
For example:
```
input = np.array([[[1, 2, 3, 4], # Input shape: (2, 3, 4)
[5, 6, 7, 8],
[9, 8, 7, 6]],
[[5, 4, 3, 2],
[1, 2, 3, 4],
[5, 6, 7, 8]]])
# A main diagonal from each batch.
tf.matrix_diag_part(input) ==> [[1, 6, 7], # Output shape: (2, 3)
[5, 2, 7]]
# A superdiagonal from each batch.
tf.matrix_diag_part(input, k = 1)
==> [[2, 7, 6], # Output shape: (2, 3)
[4, 3, 8]]
# A tridiagonal band from each batch.
tf.matrix_diag_part(input, k = (-1, 1))
==> [[[2, 7, 6], # Output shape: (2, 3, 3)
[1, 6, 7],
[5, 8, 0]],
[[4, 3, 8],
[5, 2, 7],
[1, 6, 0]]]
# Padding value = 9
tf.matrix_diag_part(input, k = (1, 3), padding_value = 9)
==> [[[4, 9, 9], # Output shape: (2, 3, 3)
[3, 8, 9],
[2, 7, 6]],
[[2, 9, 9],
[3, 4, 9],
[4, 3, 8]]]
```
Args:
input: A `Tensor`. Rank `r` tensor where `r >= 2`.
k: A `Tensor` of type `int32`.
Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main
diagonal, and negative value means subdiagonals. `k` can be a single integer
(for a single diagonal) or a pair of integers specifying the low and high ends
of a matrix band. `k[0]` must not be larger than `k[1]`.
padding_value: A `Tensor`. Must have the same type as `input`.
The value to fill the area outside the specified diagonal band with.
Default is 0.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixDiagPartV2", name,
tld.op_callbacks, input, k, padding_value)
return _result
except _core._FallbackException:
try:
return matrix_diag_part_v2_eager_fallback(
input, k, padding_value, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixDiagPartV2", input=input, k=k, padding_value=padding_value,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixDiagPartV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixDiagPartV2 = tf_export("raw_ops.MatrixDiagPartV2")(_ops.to_raw_op(matrix_diag_part_v2))
def matrix_diag_part_v2_eager_fallback(input, k, padding_value, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, padding_value], ctx)
(input, padding_value) = _inputs_T
k = _ops.convert_to_tensor(k, _dtypes.int32)
_inputs_flat = [input, k, padding_value]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"MatrixDiagPartV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixDiagPartV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_diag_part_v3(input, k, padding_value, align="RIGHT_LEFT", name=None):
r"""Returns the batched diagonal part of a batched tensor.
Returns a tensor with the `k[0]`-th to `k[1]`-th diagonals of the batched
`input`.
Assume `input` has `r` dimensions `[I, J, ..., L, M, N]`.
Let `max_diag_len` be the maximum length among all diagonals to be extracted,
`max_diag_len = min(M + min(k[1], 0), N + min(-k[0], 0))`
Let `num_diags` be the number of diagonals to extract,
`num_diags = k[1] - k[0] + 1`.
If `num_diags == 1`, the output tensor is of rank `r - 1` with shape
`[I, J, ..., L, max_diag_len]` and values:
```
diagonal[i, j, ..., l, n]
= input[i, j, ..., l, n+y, n+x] ; if 0 <= n+y < M and 0 <= n+x < N,
padding_value ; otherwise.
```
where `y = max(-k[1], 0)`, `x = max(k[1], 0)`.
Otherwise, the output tensor has rank `r` with dimensions
`[I, J, ..., L, num_diags, max_diag_len]` with values:
```
diagonal[i, j, ..., l, m, n]
= input[i, j, ..., l, n+y, n+x] ; if 0 <= n+y < M and 0 <= n+x < N,
padding_value ; otherwise.
```
where `d = k[1] - m`, `y = max(-d, 0) - offset`, and `x = max(d, 0) - offset`.
`offset` is zero except when the alignment of the diagonal is to the right.
```
offset = max_diag_len - diag_len(d) ; if (`align` in {RIGHT_LEFT, RIGHT_RIGHT}
and `d >= 0`) or
(`align` in {LEFT_RIGHT, RIGHT_RIGHT}
and `d <= 0`)
0 ; otherwise
```
where `diag_len(d) = min(cols - max(d, 0), rows + min(d, 0))`.
The input must be at least a matrix.
For example:
```
input = np.array([[[1, 2, 3, 4], # Input shape: (2, 3, 4)
[5, 6, 7, 8],
[9, 8, 7, 6]],
[[5, 4, 3, 2],
[1, 2, 3, 4],
[5, 6, 7, 8]]])
# A main diagonal from each batch.
tf.matrix_diag_part(input) ==> [[1, 6, 7], # Output shape: (2, 3)
[5, 2, 7]]
# A superdiagonal from each batch.
tf.matrix_diag_part(input, k = 1)
==> [[2, 7, 6], # Output shape: (2, 3)
[4, 3, 8]]
# A band from each batch.
tf.matrix_diag_part(input, k = (-1, 2))
==> [[[0, 3, 8], # Output shape: (2, 4, 3)
[2, 7, 6],
[1, 6, 7],
[5, 8, 0]],
[[0, 3, 4],
[4, 3, 8],
[5, 2, 7],
[1, 6, 0]]]
# LEFT_RIGHT alignment.
tf.matrix_diag_part(input, k = (-1, 2), align="LEFT_RIGHT")
==> [[[3, 8, 0], # Output shape: (2, 4, 3)
[2, 7, 6],
[1, 6, 7],
[0, 5, 8]],
[[3, 4, 0],
[4, 3, 8],
[5, 2, 7],
[0, 1, 6]]]
# max_diag_len can be shorter than the main diagonal.
tf.matrix_diag_part(input, k = (-2, -1))
==> [[[5, 8],
[9, 0]],
[[1, 6],
[5, 0]]]
# padding_value = 9
tf.matrix_diag_part(input, k = (1, 3), padding_value = 9)
==> [[[9, 9, 4], # Output shape: (2, 3, 3)
[9, 3, 8],
[2, 7, 6]],
[[9, 9, 2],
[9, 3, 4],
[4, 3, 8]]]
```
Args:
input: A `Tensor`. Rank `r` tensor where `r >= 2`.
k: A `Tensor` of type `int32`.
Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main
diagonal, and negative value means subdiagonals. `k` can be a single integer
(for a single diagonal) or a pair of integers specifying the low and high ends
of a matrix band. `k[0]` must not be larger than `k[1]`.
padding_value: A `Tensor`. Must have the same type as `input`.
The value to fill the area outside the specified diagonal band with.
Default is 0.
align: An optional `string` from: `"LEFT_RIGHT", "RIGHT_LEFT", "LEFT_LEFT", "RIGHT_RIGHT"`. Defaults to `"RIGHT_LEFT"`.
Some diagonals are shorter than `max_diag_len` and need to be padded. `align` is
a string specifying how superdiagonals and subdiagonals should be aligned,
respectively. There are four possible alignments: "RIGHT_LEFT" (default),
"LEFT_RIGHT", "LEFT_LEFT", and "RIGHT_RIGHT". "RIGHT_LEFT" aligns superdiagonals
to the right (left-pads the row) and subdiagonals to the left (right-pads the
row). It is the packing format LAPACK uses. cuSPARSE uses "LEFT_RIGHT", which is
the opposite alignment.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixDiagPartV3", name,
tld.op_callbacks, input, k, padding_value, "align", align)
return _result
except _core._FallbackException:
try:
return matrix_diag_part_v3_eager_fallback(
input, k, padding_value, align=align, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if align is None:
align = "RIGHT_LEFT"
align = _execute.make_str(align, "align")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixDiagPartV3", input=input, k=k, padding_value=padding_value,
align=align, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "align", _op.get_attr("align"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixDiagPartV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixDiagPartV3 = tf_export("raw_ops.MatrixDiagPartV3")(_ops.to_raw_op(matrix_diag_part_v3))
def matrix_diag_part_v3_eager_fallback(input, k, padding_value, align, name, ctx):
if align is None:
align = "RIGHT_LEFT"
align = _execute.make_str(align, "align")
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, padding_value], ctx)
(input, padding_value) = _inputs_T
k = _ops.convert_to_tensor(k, _dtypes.int32)
_inputs_flat = [input, k, padding_value]
_attrs = ("T", _attr_T, "align", align)
_result = _execute.execute(b"MatrixDiagPartV3", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixDiagPartV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_diag_v2(diagonal, k, num_rows, num_cols, padding_value, name=None):
r"""Returns a batched diagonal tensor with given batched diagonal values.
Returns a tensor with the contents in `diagonal` as `k[0]`-th to `k[1]`-th
diagonals of a matrix, with everything else padded with `padding`. `num_rows`
and `num_cols` specify the dimension of the innermost matrix of the output. If
both are not specified, the op assumes the innermost matrix is square and infers
its size from `k` and the innermost dimension of `diagonal`. If only one of them
is specified, the op assumes the unspecified value is the smallest possible
based on other criteria.
Let `diagonal` have `r` dimensions `[I, J, ..., L, M, N]`. The output tensor has
rank `r+1` with shape `[I, J, ..., L, M, num_rows, num_cols]` when only one
diagonal is given (`k` is an integer or `k[0] == k[1]`). Otherwise, it has rank
`r` with shape `[I, J, ..., L, num_rows, num_cols]`.
The second innermost dimension of `diagonal` has double meaning.
When `k` is scalar or `k[0] == k[1]`, `M` is part of the batch size
[I, J, ..., M], and the output tensor is:
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, n-max(d_upper, 0)] ; if n - m == d_upper
padding_value ; otherwise
```
Otherwise, `M` is treated as the number of diagonals for the matrix in the
same batch (`M = k[1]-k[0]+1`), and the output tensor is:
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, diag_index, index_in_diag] ; if k[0] <= d <= k[1]
padding_value ; otherwise
```
where `d = n - m`, `diag_index = k[1] - d`, and `index_in_diag = n - max(d, 0)`.
For example:
```
# The main diagonal.
diagonal = np.array([[1, 2, 3, 4], # Input shape: (2, 4)
[5, 6, 7, 8]])
tf.matrix_diag(diagonal) ==> [[[1, 0, 0, 0], # Output shape: (2, 4, 4)
[0, 2, 0, 0],
[0, 0, 3, 0],
[0, 0, 0, 4]],
[[5, 0, 0, 0],
[0, 6, 0, 0],
[0, 0, 7, 0],
[0, 0, 0, 8]]]
# A superdiagonal (per batch).
diagonal = np.array([[1, 2, 3], # Input shape: (2, 3)
[4, 5, 6]])
tf.matrix_diag(diagonal, k = 1)
==> [[[0, 1, 0, 0], # Output shape: (2, 4, 4)
[0, 0, 2, 0],
[0, 0, 0, 3],
[0, 0, 0, 0]],
[[0, 4, 0, 0],
[0, 0, 5, 0],
[0, 0, 0, 6],
[0, 0, 0, 0]]]
# A band of diagonals.
diagonals = np.array([[[1, 2, 3], # Input shape: (2, 2, 3)
[4, 5, 0]],
[[6, 7, 9],
[9, 1, 0]]])
tf.matrix_diag(diagonals, k = (-1, 0))
==> [[[1, 0, 0], # Output shape: (2, 3, 3)
[4, 2, 0],
[0, 5, 3]],
[[6, 0, 0],
[9, 7, 0],
[0, 1, 9]]]
# Rectangular matrix.
diagonal = np.array([1, 2]) # Input shape: (2)
tf.matrix_diag(diagonal, k = -1, num_rows = 3, num_cols = 4)
==> [[0, 0, 0, 0], # Output shape: (3, 4)
[1, 0, 0, 0],
[0, 2, 0, 0]]
# Rectangular matrix with inferred num_cols and padding_value = 9.
tf.matrix_diag(diagonal, k = -1, num_rows = 3, padding_value = 9)
==> [[9, 9], # Output shape: (3, 2)
[1, 9],
[9, 2]]
```
Args:
diagonal: A `Tensor`. Rank `r`, where `r >= 1`
k: A `Tensor` of type `int32`.
Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main
diagonal, and negative value means subdiagonals. `k` can be a single integer
(for a single diagonal) or a pair of integers specifying the low and high ends
of a matrix band. `k[0]` must not be larger than `k[1]`.
num_rows: A `Tensor` of type `int32`.
The number of rows of the output matrix. If it is not provided, the op assumes
the output matrix is a square matrix and infers the matrix size from k and the
innermost dimension of `diagonal`.
num_cols: A `Tensor` of type `int32`.
The number of columns of the output matrix. If it is not provided, the op
assumes the output matrix is a square matrix and infers the matrix size from
k and the innermost dimension of `diagonal`.
padding_value: A `Tensor`. Must have the same type as `diagonal`.
The number to fill the area outside the specified diagonal band with.
Default is 0.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `diagonal`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixDiagV2", name,
tld.op_callbacks, diagonal, k, num_rows, num_cols, padding_value)
return _result
except _core._FallbackException:
try:
return matrix_diag_v2_eager_fallback(
diagonal, k, num_rows, num_cols, padding_value, name=name,
ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixDiagV2", diagonal=diagonal, k=k, num_rows=num_rows,
num_cols=num_cols, padding_value=padding_value,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixDiagV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixDiagV2 = tf_export("raw_ops.MatrixDiagV2")(_ops.to_raw_op(matrix_diag_v2))
def matrix_diag_v2_eager_fallback(diagonal, k, num_rows, num_cols, padding_value, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([diagonal, padding_value], ctx)
(diagonal, padding_value) = _inputs_T
k = _ops.convert_to_tensor(k, _dtypes.int32)
num_rows = _ops.convert_to_tensor(num_rows, _dtypes.int32)
num_cols = _ops.convert_to_tensor(num_cols, _dtypes.int32)
_inputs_flat = [diagonal, k, num_rows, num_cols, padding_value]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"MatrixDiagV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixDiagV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_diag_v3(diagonal, k, num_rows, num_cols, padding_value, align="RIGHT_LEFT", name=None):
r"""Returns a batched diagonal tensor with given batched diagonal values.
Returns a tensor with the contents in `diagonal` as `k[0]`-th to `k[1]`-th
diagonals of a matrix, with everything else padded with `padding`. `num_rows`
and `num_cols` specify the dimension of the innermost matrix of the output. If
both are not specified, the op assumes the innermost matrix is square and infers
its size from `k` and the innermost dimension of `diagonal`. If only one of them
is specified, the op assumes the unspecified value is the smallest possible
based on other criteria.
Let `diagonal` have `r` dimensions `[I, J, ..., L, M, N]`. The output tensor has
rank `r+1` with shape `[I, J, ..., L, M, num_rows, num_cols]` when only one
diagonal is given (`k` is an integer or `k[0] == k[1]`). Otherwise, it has rank
`r` with shape `[I, J, ..., L, num_rows, num_cols]`.
The second innermost dimension of `diagonal` has double meaning.
When `k` is scalar or `k[0] == k[1]`, `M` is part of the batch size
[I, J, ..., M], and the output tensor is:
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, n-max(d_upper, 0)] ; if n - m == d_upper
padding_value ; otherwise
```
Otherwise, `M` is treated as the number of diagonals for the matrix in the
same batch (`M = k[1]-k[0]+1`), and the output tensor is:
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, diag_index, index_in_diag] ; if k[0] <= d <= k[1]
padding_value ; otherwise
```
where `d = n - m`, `diag_index = [k] - d`, and
`index_in_diag = n - max(d, 0) + offset`.
`offset` is zero except when the alignment of the diagonal is to the right.
```
offset = max_diag_len - diag_len(d) ; if (`align` in {RIGHT_LEFT, RIGHT_RIGHT}
and `d >= 0`) or
(`align` in {LEFT_RIGHT, RIGHT_RIGHT}
and `d <= 0`)
0 ; otherwise
```
where `diag_len(d) = min(cols - max(d, 0), rows + min(d, 0))`.
For example:
```
# The main diagonal.
diagonal = np.array([[1, 2, 3, 4], # Input shape: (2, 4)
[5, 6, 7, 8]])
tf.matrix_diag(diagonal) ==> [[[1, 0, 0, 0], # Output shape: (2, 4, 4)
[0, 2, 0, 0],
[0, 0, 3, 0],
[0, 0, 0, 4]],
[[5, 0, 0, 0],
[0, 6, 0, 0],
[0, 0, 7, 0],
[0, 0, 0, 8]]]
# A superdiagonal (per batch).
diagonal = np.array([[1, 2, 3], # Input shape: (2, 3)
[4, 5, 6]])
tf.matrix_diag(diagonal, k = 1)
==> [[[0, 1, 0, 0], # Output shape: (2, 4, 4)
[0, 0, 2, 0],
[0, 0, 0, 3],
[0, 0, 0, 0]],
[[0, 4, 0, 0],
[0, 0, 5, 0],
[0, 0, 0, 6],
[0, 0, 0, 0]]]
# A tridiagonal band (per batch).
diagonals = np.array([[[0, 8, 9], # Input shape: (2, 2, 3)
[1, 2, 3],
[4, 5, 0]],
[[0, 2, 3],
[6, 7, 9],
[9, 1, 0]]])
tf.matrix_diag(diagonals, k = (-1, 1))
==> [[[1, 8, 0], # Output shape: (2, 3, 3)
[4, 2, 9],
[0, 5, 3]],
[[6, 2, 0],
[9, 7, 3],
[0, 1, 9]]]
# LEFT_RIGHT alignment.
diagonals = np.array([[[8, 9, 0], # Input shape: (2, 2, 3)
[1, 2, 3],
[0, 4, 5]],
[[2, 3, 0],
[6, 7, 9],
[0, 9, 1]]])
tf.matrix_diag(diagonals, k = (-1, 1), align="LEFT_RIGHT")
==> [[[1, 8, 0], # Output shape: (2, 3, 3)
[4, 2, 9],
[0, 5, 3]],
[[6, 2, 0],
[9, 7, 3],
[0, 1, 9]]]
# Rectangular matrix.
diagonal = np.array([1, 2]) # Input shape: (2)
tf.matrix_diag(diagonal, k = -1, num_rows = 3, num_cols = 4)
==> [[0, 0, 0, 0], # Output shape: (3, 4)
[1, 0, 0, 0],
[0, 2, 0, 0]]
# Rectangular matrix with inferred num_cols and padding_value = 9.
tf.matrix_diag(diagonal, k = -1, num_rows = 3, padding_value = 9)
==> [[9, 9], # Output shape: (3, 2)
[1, 9],
[9, 2]]
```
Args:
diagonal: A `Tensor`. Rank `r`, where `r >= 1`
k: A `Tensor` of type `int32`.
Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main
diagonal, and negative value means subdiagonals. `k` can be a single integer
(for a single diagonal) or a pair of integers specifying the low and high ends
of a matrix band. `k[0]` must not be larger than `k[1]`.
num_rows: A `Tensor` of type `int32`.
The number of rows of the output matrix. If it is not provided, the op assumes
the output matrix is a square matrix and infers the matrix size from k and the
innermost dimension of `diagonal`.
num_cols: A `Tensor` of type `int32`.
The number of columns of the output matrix. If it is not provided, the op
assumes the output matrix is a square matrix and infers the matrix size from
k and the innermost dimension of `diagonal`.
padding_value: A `Tensor`. Must have the same type as `diagonal`.
The number to fill the area outside the specified diagonal band with.
Default is 0.
align: An optional `string` from: `"LEFT_RIGHT", "RIGHT_LEFT", "LEFT_LEFT", "RIGHT_RIGHT"`. Defaults to `"RIGHT_LEFT"`.
Some diagonals are shorter than `max_diag_len` and need to be padded. `align` is
a string specifying how superdiagonals and subdiagonals should be aligned,
respectively. There are four possible alignments: "RIGHT_LEFT" (default),
"LEFT_RIGHT", "LEFT_LEFT", and "RIGHT_RIGHT". "RIGHT_LEFT" aligns superdiagonals
to the right (left-pads the row) and subdiagonals to the left (right-pads the
row). It is the packing format LAPACK uses. cuSPARSE uses "LEFT_RIGHT", which is
the opposite alignment.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `diagonal`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixDiagV3", name,
tld.op_callbacks, diagonal, k, num_rows, num_cols, padding_value,
"align", align)
return _result
except _core._FallbackException:
try:
return matrix_diag_v3_eager_fallback(
diagonal, k, num_rows, num_cols, padding_value, align=align,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if align is None:
align = "RIGHT_LEFT"
align = _execute.make_str(align, "align")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixDiagV3", diagonal=diagonal, k=k, num_rows=num_rows,
num_cols=num_cols, padding_value=padding_value,
align=align, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "align", _op.get_attr("align"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixDiagV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixDiagV3 = tf_export("raw_ops.MatrixDiagV3")(_ops.to_raw_op(matrix_diag_v3))
def matrix_diag_v3_eager_fallback(diagonal, k, num_rows, num_cols, padding_value, align, name, ctx):
if align is None:
align = "RIGHT_LEFT"
align = _execute.make_str(align, "align")
_attr_T, _inputs_T = _execute.args_to_matching_eager([diagonal, padding_value], ctx)
(diagonal, padding_value) = _inputs_T
k = _ops.convert_to_tensor(k, _dtypes.int32)
num_rows = _ops.convert_to_tensor(num_rows, _dtypes.int32)
num_cols = _ops.convert_to_tensor(num_cols, _dtypes.int32)
_inputs_flat = [diagonal, k, num_rows, num_cols, padding_value]
_attrs = ("T", _attr_T, "align", align)
_result = _execute.execute(b"MatrixDiagV3", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixDiagV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_set_diag(input, diagonal, name=None):
r"""Returns a batched matrix tensor with new batched diagonal values.
Given `input` and `diagonal`, this operation returns a tensor with the
same shape and values as `input`, except for the main diagonal of the
innermost matrices. These will be overwritten by the values in `diagonal`.
The output is computed as follows:
Assume `input` has `k+1` dimensions `[I, J, K, ..., M, N]` and `diagonal` has
`k` dimensions `[I, J, K, ..., min(M, N)]`. Then the output is a
tensor of rank `k+1` with dimensions `[I, J, K, ..., M, N]` where:
* `output[i, j, k, ..., m, n] = diagonal[i, j, k, ..., n]` for `m == n`.
* `output[i, j, k, ..., m, n] = input[i, j, k, ..., m, n]` for `m != n`.
Args:
input: A `Tensor`. Rank `k+1`, where `k >= 1`.
diagonal: A `Tensor`. Must have the same type as `input`.
Rank `k`, where `k >= 1`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixSetDiag", name,
tld.op_callbacks, input, diagonal)
return _result
except _core._FallbackException:
try:
return matrix_set_diag_eager_fallback(
input, diagonal, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixSetDiag", input=input, diagonal=diagonal, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixSetDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixSetDiag = tf_export("raw_ops.MatrixSetDiag")(_ops.to_raw_op(matrix_set_diag))
def matrix_set_diag_eager_fallback(input, diagonal, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, diagonal], ctx)
(input, diagonal) = _inputs_T
_inputs_flat = [input, diagonal]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"MatrixSetDiag", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixSetDiag", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_set_diag_v2(input, diagonal, k, name=None):
r"""Returns a batched matrix tensor with new batched diagonal values.
Given `input` and `diagonal`, this operation returns a tensor with the
same shape and values as `input`, except for the specified diagonals of the
innermost matrices. These will be overwritten by the values in `diagonal`.
`input` has `r+1` dimensions `[I, J, ..., L, M, N]`. When `k` is scalar or
`k[0] == k[1]`, `diagonal` has `r` dimensions `[I, J, ..., L, max_diag_len]`.
Otherwise, it has `r+1` dimensions `[I, J, ..., L, num_diags, max_diag_len]`.
`num_diags` is the number of diagonals, `num_diags = k[1] - k[0] + 1`.
`max_diag_len` is the longest diagonal in the range `[k[0], k[1]]`,
`max_diag_len = min(M + min(k[1], 0), N + min(-k[0], 0))`
The output is a tensor of rank `k+1` with dimensions `[I, J, ..., L, M, N]`.
If `k` is scalar or `k[0] == k[1]`:
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, n-max(k[1], 0)] ; if n - m == k[1]
input[i, j, ..., l, m, n] ; otherwise
```
Otherwise,
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, diag_index, index_in_diag] ; if k[0] <= d <= k[1]
input[i, j, ..., l, m, n] ; otherwise
```
where `d = n - m`, `diag_index = k[1] - d`, and `index_in_diag = n - max(d, 0)`.
For example:
```
# The main diagonal.
input = np.array([[[7, 7, 7, 7], # Input shape: (2, 3, 4)
[7, 7, 7, 7],
[7, 7, 7, 7]],
[[7, 7, 7, 7],
[7, 7, 7, 7],
[7, 7, 7, 7]]])
diagonal = np.array([[1, 2, 3], # Diagonal shape: (2, 3)
[4, 5, 6]])
tf.matrix_set_diag(diagonal) ==> [[[1, 7, 7, 7], # Output shape: (2, 3, 4)
[7, 2, 7, 7],
[7, 7, 3, 7]],
[[4, 7, 7, 7],
[7, 5, 7, 7],
[7, 7, 6, 7]]]
# A superdiagonal (per batch).
tf.matrix_set_diag(diagonal, k = 1)
==> [[[7, 1, 7, 7], # Output shape: (2, 3, 4)
[7, 7, 2, 7],
[7, 7, 7, 3]],
[[7, 4, 7, 7],
[7, 7, 5, 7],
[7, 7, 7, 6]]]
# A band of diagonals.
diagonals = np.array([[[1, 2, 3], # Diagonal shape: (2, 2, 3)
[4, 5, 0]],
[[6, 1, 2],
[3, 4, 0]]])
tf.matrix_set_diag(diagonals, k = (-1, 0))
==> [[[1, 7, 7, 7], # Output shape: (2, 3, 4)
[4, 2, 7, 7],
[0, 5, 3, 7]],
[[6, 7, 7, 7],
[3, 1, 7, 7],
[7, 4, 2, 7]]]
```
Args:
input: A `Tensor`. Rank `r+1`, where `r >= 1`.
diagonal: A `Tensor`. Must have the same type as `input`.
Rank `r` when `k` is an integer or `k[0] == k[1]`. Otherwise, it has rank `r+1`.
`k >= 1`.
k: A `Tensor` of type `int32`.
Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main
diagonal, and negative value means subdiagonals. `k` can be a single integer
(for a single diagonal) or a pair of integers specifying the low and high ends
of a matrix band. `k[0]` must not be larger than `k[1]`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixSetDiagV2", name,
tld.op_callbacks, input, diagonal, k)
return _result
except _core._FallbackException:
try:
return matrix_set_diag_v2_eager_fallback(
input, diagonal, k, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixSetDiagV2", input=input, diagonal=diagonal, k=k, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixSetDiagV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixSetDiagV2 = tf_export("raw_ops.MatrixSetDiagV2")(_ops.to_raw_op(matrix_set_diag_v2))
def matrix_set_diag_v2_eager_fallback(input, diagonal, k, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, diagonal], ctx)
(input, diagonal) = _inputs_T
k = _ops.convert_to_tensor(k, _dtypes.int32)
_inputs_flat = [input, diagonal, k]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"MatrixSetDiagV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixSetDiagV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def matrix_set_diag_v3(input, diagonal, k, align="RIGHT_LEFT", name=None):
r"""Returns a batched matrix tensor with new batched diagonal values.
Given `input` and `diagonal`, this operation returns a tensor with the
same shape and values as `input`, except for the specified diagonals of the
innermost matrices. These will be overwritten by the values in `diagonal`.
`input` has `r+1` dimensions `[I, J, ..., L, M, N]`. When `k` is scalar or
`k[0] == k[1]`, `diagonal` has `r` dimensions `[I, J, ..., L, max_diag_len]`.
Otherwise, it has `r+1` dimensions `[I, J, ..., L, num_diags, max_diag_len]`.
`num_diags` is the number of diagonals, `num_diags = k[1] - k[0] + 1`.
`max_diag_len` is the longest diagonal in the range `[k[0], k[1]]`,
`max_diag_len = min(M + min(k[1], 0), N + min(-k[0], 0))`
The output is a tensor of rank `k+1` with dimensions `[I, J, ..., L, M, N]`.
If `k` is scalar or `k[0] == k[1]`:
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, n-max(k[1], 0)] ; if n - m == k[1]
input[i, j, ..., l, m, n] ; otherwise
```
Otherwise,
```
output[i, j, ..., l, m, n]
= diagonal[i, j, ..., l, diag_index, index_in_diag] ; if k[0] <= d <= k[1]
input[i, j, ..., l, m, n] ; otherwise
```
where `d = n - m`, `diag_index = k[1] - d`, and
`index_in_diag = n - max(d, 0) + offset`.
`offset` is zero except when the alignment of the diagonal is to the right.
```
offset = max_diag_len - diag_len(d) ; if (`align` in {RIGHT_LEFT, RIGHT_RIGHT}
and `d >= 0`) or
(`align` in {LEFT_RIGHT, RIGHT_RIGHT}
and `d <= 0`)
0 ; otherwise
```
where `diag_len(d) = min(cols - max(d, 0), rows + min(d, 0))`.
For example:
```
# The main diagonal.
input = np.array([[[7, 7, 7, 7], # Input shape: (2, 3, 4)
[7, 7, 7, 7],
[7, 7, 7, 7]],
[[7, 7, 7, 7],
[7, 7, 7, 7],
[7, 7, 7, 7]]])
diagonal = np.array([[1, 2, 3], # Diagonal shape: (2, 3)
[4, 5, 6]])
tf.matrix_set_diag(input, diagonal)
==> [[[1, 7, 7, 7], # Output shape: (2, 3, 4)
[7, 2, 7, 7],
[7, 7, 3, 7]],
[[4, 7, 7, 7],
[7, 5, 7, 7],
[7, 7, 6, 7]]]
# A superdiagonal (per batch).
tf.matrix_set_diag(input, diagonal, k = 1)
==> [[[7, 1, 7, 7], # Output shape: (2, 3, 4)
[7, 7, 2, 7],
[7, 7, 7, 3]],
[[7, 4, 7, 7],
[7, 7, 5, 7],
[7, 7, 7, 6]]]
# A band of diagonals.
diagonals = np.array([[[0, 9, 1], # Diagonal shape: (2, 4, 3)
[6, 5, 8],
[1, 2, 3],
[4, 5, 0]],
[[0, 1, 2],
[5, 6, 4],
[6, 1, 2],
[3, 4, 0]]])
tf.matrix_set_diag(input, diagonals, k = (-1, 2))
==> [[[1, 6, 9, 7], # Output shape: (2, 3, 4)
[4, 2, 5, 1],
[7, 5, 3, 8]],
[[6, 5, 1, 7],
[3, 1, 6, 2],
[7, 4, 2, 4]]]
# LEFT_RIGHT alignment.
diagonals = np.array([[[9, 1, 0], # Diagonal shape: (2, 4, 3)
[6, 5, 8],
[1, 2, 3],
[0, 4, 5]],
[[1, 2, 0],
[5, 6, 4],
[6, 1, 2],
[0, 3, 4]]])
tf.matrix_set_diag(input, diagonals, k = (-1, 2), align="LEFT_RIGHT")
==> [[[1, 6, 9, 7], # Output shape: (2, 3, 4)
[4, 2, 5, 1],
[7, 5, 3, 8]],
[[6, 5, 1, 7],
[3, 1, 6, 2],
[7, 4, 2, 4]]]
```
Args:
input: A `Tensor`. Rank `r+1`, where `r >= 1`.
diagonal: A `Tensor`. Must have the same type as `input`.
Rank `r` when `k` is an integer or `k[0] == k[1]`. Otherwise, it has rank `r+1`.
`k >= 1`.
k: A `Tensor` of type `int32`.
Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main
diagonal, and negative value means subdiagonals. `k` can be a single integer
(for a single diagonal) or a pair of integers specifying the low and high ends
of a matrix band. `k[0]` must not be larger than `k[1]`.
align: An optional `string` from: `"LEFT_RIGHT", "RIGHT_LEFT", "LEFT_LEFT", "RIGHT_RIGHT"`. Defaults to `"RIGHT_LEFT"`.
Some diagonals are shorter than `max_diag_len` and need to be padded. `align` is
a string specifying how superdiagonals and subdiagonals should be aligned,
respectively. There are four possible alignments: "RIGHT_LEFT" (default),
"LEFT_RIGHT", "LEFT_LEFT", and "RIGHT_RIGHT". "RIGHT_LEFT" aligns superdiagonals
to the right (left-pads the row) and subdiagonals to the left (right-pads the
row). It is the packing format LAPACK uses. cuSPARSE uses "LEFT_RIGHT", which is
the opposite alignment.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MatrixSetDiagV3", name,
tld.op_callbacks, input, diagonal, k, "align", align)
return _result
except _core._FallbackException:
try:
return matrix_set_diag_v3_eager_fallback(
input, diagonal, k, align=align, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if align is None:
align = "RIGHT_LEFT"
align = _execute.make_str(align, "align")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MatrixSetDiagV3", input=input, diagonal=diagonal, k=k, align=align,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "align", _op.get_attr("align"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MatrixSetDiagV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MatrixSetDiagV3 = tf_export("raw_ops.MatrixSetDiagV3")(_ops.to_raw_op(matrix_set_diag_v3))
def matrix_set_diag_v3_eager_fallback(input, diagonal, k, align, name, ctx):
if align is None:
align = "RIGHT_LEFT"
align = _execute.make_str(align, "align")
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, diagonal], ctx)
(input, diagonal) = _inputs_T
k = _ops.convert_to_tensor(k, _dtypes.int32)
_inputs_flat = [input, diagonal, k]
_attrs = ("T", _attr_T, "align", align)
_result = _execute.execute(b"MatrixSetDiagV3", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MatrixSetDiagV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def mirror_pad(input, paddings, mode, name=None):
r"""Pads a tensor with mirrored values.
This operation pads a `input` with mirrored values according to the `paddings`
you specify. `paddings` is an integer tensor with shape `[n, 2]`, where n is
the rank of `input`. For each dimension D of `input`, `paddings[D, 0]` indicates
how many values to add before the contents of `input` in that dimension, and
`paddings[D, 1]` indicates how many values to add after the contents of `input`
in that dimension. Both `paddings[D, 0]` and `paddings[D, 1]` must be no greater
than `input.dim_size(D)` (or `input.dim_size(D) - 1`) if `copy_border` is true
(if false, respectively).
The padded size of each dimension D of the output is:
`paddings(D, 0) + input.dim_size(D) + paddings(D, 1)`
For example:
```
# 't' is [[1, 2, 3], [4, 5, 6]].
# 'paddings' is [[1, 1]], [2, 2]].
# 'mode' is SYMMETRIC.
# rank of 't' is 2.
pad(t, paddings) ==> [[2, 1, 1, 2, 3, 3, 2]
[2, 1, 1, 2, 3, 3, 2]
[5, 4, 4, 5, 6, 6, 5]
[5, 4, 4, 5, 6, 6, 5]]
```
Args:
input: A `Tensor`. The input tensor to be padded.
paddings: A `Tensor`. Must be one of the following types: `int32`, `int64`.
A two-column matrix specifying the padding sizes. The number of
rows must be the same as the rank of `input`.
mode: A `string` from: `"REFLECT", "SYMMETRIC"`.
Either `REFLECT` or `SYMMETRIC`. In reflect mode the padded regions
do not include the borders, while in symmetric mode the padded regions
do include the borders. For example, if `input` is `[1, 2, 3]` and `paddings`
is `[0, 2]`, then the output is `[1, 2, 3, 2, 1]` in reflect mode, and
it is `[1, 2, 3, 3, 2]` in symmetric mode.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MirrorPad", name,
tld.op_callbacks, input, paddings, "mode", mode)
return _result
except _core._FallbackException:
try:
return mirror_pad_eager_fallback(
input, paddings, mode=mode, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
mode = _execute.make_str(mode, "mode")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MirrorPad", input=input, paddings=paddings, mode=mode, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tpaddings",
_op._get_attr_type("Tpaddings"), "mode", _op.get_attr("mode"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MirrorPad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MirrorPad = tf_export("raw_ops.MirrorPad")(_ops.to_raw_op(mirror_pad))
def mirror_pad_eager_fallback(input, paddings, mode, name, ctx):
mode = _execute.make_str(mode, "mode")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tpaddings, (paddings,) = _execute.args_to_matching_eager([paddings], ctx, _dtypes.int32)
_inputs_flat = [input, paddings]
_attrs = ("T", _attr_T, "Tpaddings", _attr_Tpaddings, "mode", mode)
_result = _execute.execute(b"MirrorPad", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MirrorPad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def mirror_pad_grad(input, paddings, mode, name=None):
r"""Gradient op for `MirrorPad` op. This op folds a mirror-padded tensor.
This operation folds the padded areas of `input` by `MirrorPad` according to the
`paddings` you specify. `paddings` must be the same as `paddings` argument
given to the corresponding `MirrorPad` op.
The folded size of each dimension D of the output is:
`input.dim_size(D) - paddings(D, 0) - paddings(D, 1)`
For example:
```
# 't' is [[1, 2, 3], [4, 5, 6], [7, 8, 9]].
# 'paddings' is [[0, 1]], [0, 1]].
# 'mode' is SYMMETRIC.
# rank of 't' is 2.
pad(t, paddings) ==> [[ 1, 5]
[11, 28]]
```
Args:
input: A `Tensor`. The input tensor to be folded.
paddings: A `Tensor`. Must be one of the following types: `int32`, `int64`.
A two-column matrix specifying the padding sizes. The number of
rows must be the same as the rank of `input`.
mode: A `string` from: `"REFLECT", "SYMMETRIC"`.
The mode used in the `MirrorPad` op.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "MirrorPadGrad", name,
tld.op_callbacks, input, paddings, "mode", mode)
return _result
except _core._FallbackException:
try:
return mirror_pad_grad_eager_fallback(
input, paddings, mode=mode, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
mode = _execute.make_str(mode, "mode")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"MirrorPadGrad", input=input, paddings=paddings, mode=mode, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tpaddings",
_op._get_attr_type("Tpaddings"), "mode", _op.get_attr("mode"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"MirrorPadGrad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
MirrorPadGrad = tf_export("raw_ops.MirrorPadGrad")(_ops.to_raw_op(mirror_pad_grad))
def mirror_pad_grad_eager_fallback(input, paddings, mode, name, ctx):
mode = _execute.make_str(mode, "mode")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tpaddings, (paddings,) = _execute.args_to_matching_eager([paddings], ctx, _dtypes.int32)
_inputs_flat = [input, paddings]
_attrs = ("T", _attr_T, "Tpaddings", _attr_Tpaddings, "mode", mode)
_result = _execute.execute(b"MirrorPadGrad", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"MirrorPadGrad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def one_hot(indices, depth, on_value, off_value, axis=-1, name=None):
r"""Returns a one-hot tensor.
The locations represented by indices in `indices` take value `on_value`,
while all other locations take value `off_value`.
If the input `indices` is rank `N`, the output will have rank `N+1`,
The new axis is created at dimension `axis` (default: the new axis is
appended at the end).
If `indices` is a scalar the output shape will be a vector of length `depth`.
If `indices` is a vector of length `features`, the output shape will be:
```
features x depth if axis == -1
depth x features if axis == 0
```
If `indices` is a matrix (batch) with shape `[batch, features]`,
the output shape will be:
```
batch x features x depth if axis == -1
batch x depth x features if axis == 1
depth x batch x features if axis == 0
```
Examples
=========
Suppose that
```
indices = [0, 2, -1, 1]
depth = 3
on_value = 5.0
off_value = 0.0
axis = -1
```
Then output is `[4 x 3]`:
```
output =
[5.0 0.0 0.0] // one_hot(0)
[0.0 0.0 5.0] // one_hot(2)
[0.0 0.0 0.0] // one_hot(-1)
[0.0 5.0 0.0] // one_hot(1)
```
Suppose that
```
indices = [0, 2, -1, 1]
depth = 3
on_value = 0.0
off_value = 3.0
axis = 0
```
Then output is `[3 x 4]`:
```
output =
[0.0 3.0 3.0 3.0]
[3.0 3.0 3.0 0.0]
[3.0 3.0 3.0 3.0]
[3.0 0.0 3.0 3.0]
// ^ one_hot(0)
// ^ one_hot(2)
// ^ one_hot(-1)
// ^ one_hot(1)
```
Suppose that
```
indices = [[0, 2], [1, -1]]
depth = 3
on_value = 1.0
off_value = 0.0
axis = -1
```
Then output is `[2 x 2 x 3]`:
```
output =
[
[1.0, 0.0, 0.0] // one_hot(0)
[0.0, 0.0, 1.0] // one_hot(2)
][
[0.0, 1.0, 0.0] // one_hot(1)
[0.0, 0.0, 0.0] // one_hot(-1)
]
```
Args:
indices: A `Tensor`. Must be one of the following types: `uint8`, `int32`, `int64`.
A tensor of indices.
depth: A `Tensor` of type `int32`.
A scalar defining the depth of the one hot dimension.
on_value: A `Tensor`.
A scalar defining the value to fill in output when `indices[j] = i`.
off_value: A `Tensor`. Must have the same type as `on_value`.
A scalar defining the value to fill in output when `indices[j] != i`.
axis: An optional `int`. Defaults to `-1`.
The axis to fill (default: -1, a new inner-most axis).
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `on_value`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "OneHot", name,
tld.op_callbacks, indices, depth, on_value, off_value, "axis", axis)
return _result
except _core._FallbackException:
try:
return one_hot_eager_fallback(
indices, depth, on_value, off_value, axis=axis, name=name,
ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"OneHot", indices=indices, depth=depth, on_value=on_value,
off_value=off_value, axis=axis, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("axis", _op._get_attr_int("axis"), "T", _op._get_attr_type("T"),
"TI", _op._get_attr_type("TI"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"OneHot", _inputs_flat, _attrs, _result)
_result, = _result
return _result
OneHot = tf_export("raw_ops.OneHot")(_ops.to_raw_op(one_hot))
def one_hot_eager_fallback(indices, depth, on_value, off_value, axis, name, ctx):
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
_attr_T, _inputs_T = _execute.args_to_matching_eager([on_value, off_value], ctx)
(on_value, off_value) = _inputs_T
_attr_TI, (indices,) = _execute.args_to_matching_eager([indices], ctx, _dtypes.int64)
depth = _ops.convert_to_tensor(depth, _dtypes.int32)
_inputs_flat = [indices, depth, on_value, off_value]
_attrs = ("axis", axis, "T", _attr_T, "TI", _attr_TI)
_result = _execute.execute(b"OneHot", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"OneHot", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def ones_like(x, name=None):
r"""Returns a tensor of ones with the same shape and type as x.
Args:
x: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`, `int8`, `uint8`, `int16`, `uint16`, `int32`, `int64`, `complex64`, `complex128`, `bool`.
a tensor of type T.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "OnesLike", name,
tld.op_callbacks, x)
return _result
except _core._FallbackException:
try:
return ones_like_eager_fallback(
x, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"OnesLike", x=x, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"OnesLike", _inputs_flat, _attrs, _result)
_result, = _result
return _result
OnesLike = tf_export("raw_ops.OnesLike")(_ops.to_raw_op(ones_like))
def ones_like_eager_fallback(x, name, ctx):
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_inputs_flat = [x]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"OnesLike", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"OnesLike", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def pack(values, axis=0, name=None):
r"""Packs a list of `N` rank-`R` tensors into one rank-`(R+1)` tensor.
Packs the `N` tensors in `values` into a tensor with rank one higher than each
tensor in `values`, by packing them along the `axis` dimension.
Given a list of tensors of shape `(A, B, C)`;
if `axis == 0` then the `output` tensor will have the shape `(N, A, B, C)`.
if `axis == 1` then the `output` tensor will have the shape `(A, N, B, C)`.
Etc.
For example:
```
# 'x' is [1, 4]
# 'y' is [2, 5]
# 'z' is [3, 6]
pack([x, y, z]) => [[1, 4], [2, 5], [3, 6]] # Pack along first dim.
pack([x, y, z], axis=1) => [[1, 2, 3], [4, 5, 6]]
```
This is the opposite of `unpack`.
Args:
values: A list of at least 1 `Tensor` objects with the same type.
Must be of same shape and type.
axis: An optional `int`. Defaults to `0`.
Dimension along which to pack. Negative values wrap around, so the
valid range is `[-(R+1), R+1)`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `values`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Pack", name, tld.op_callbacks,
values, "axis", axis)
return _result
except _core._FallbackException:
try:
return pack_eager_fallback(
values, axis=axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'pack' Op, not %r." % values)
_attr_N = len(values)
if axis is None:
axis = 0
axis = _execute.make_int(axis, "axis")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Pack", values=values, axis=axis, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"), "T", _op._get_attr_type("T"),
"axis", _op._get_attr_int("axis"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Pack", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Pack = tf_export("raw_ops.Pack")(_ops.to_raw_op(pack))
def pack_eager_fallback(values, axis, name, ctx):
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'pack' Op, not %r." % values)
_attr_N = len(values)
if axis is None:
axis = 0
axis = _execute.make_int(axis, "axis")
_attr_T, values = _execute.args_to_matching_eager(list(values), ctx)
_inputs_flat = list(values)
_attrs = ("N", _attr_N, "T", _attr_T, "axis", axis)
_result = _execute.execute(b"Pack", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Pack", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def pad(input, paddings, name=None):
r"""Pads a tensor with zeros.
This operation pads a `input` with zeros according to the `paddings` you
specify. `paddings` is an integer tensor with shape `[Dn, 2]`, where n is the
rank of `input`. For each dimension D of `input`, `paddings[D, 0]` indicates
how many zeros to add before the contents of `input` in that dimension, and
`paddings[D, 1]` indicates how many zeros to add after the contents of `input`
in that dimension.
The padded size of each dimension D of the output is:
`paddings(D, 0) + input.dim_size(D) + paddings(D, 1)`
For example:
```
# 't' is [[1, 1], [2, 2]]
# 'paddings' is [[1, 1], [2, 2]]
# rank of 't' is 2
pad(t, paddings) ==> [[0, 0, 0, 0, 0, 0]
[0, 0, 1, 1, 0, 0]
[0, 0, 2, 2, 0, 0]
[0, 0, 0, 0, 0, 0]]
```
Args:
input: A `Tensor`.
paddings: A `Tensor`. Must be one of the following types: `int32`, `int64`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Pad", name, tld.op_callbacks,
input, paddings)
return _result
except _core._FallbackException:
try:
return pad_eager_fallback(
input, paddings, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Pad", input=input, paddings=paddings, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tpaddings",
_op._get_attr_type("Tpaddings"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Pad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Pad = tf_export("raw_ops.Pad")(_ops.to_raw_op(pad))
def pad_eager_fallback(input, paddings, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tpaddings, (paddings,) = _execute.args_to_matching_eager([paddings], ctx, _dtypes.int32)
_inputs_flat = [input, paddings]
_attrs = ("T", _attr_T, "Tpaddings", _attr_Tpaddings)
_result = _execute.execute(b"Pad", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Pad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def pad_v2(input, paddings, constant_values, name=None):
r"""Pads a tensor.
This operation pads `input` according to the `paddings` and `constant_values`
you specify. `paddings` is an integer tensor with shape `[Dn, 2]`, where n is
the rank of `input`. For each dimension D of `input`, `paddings[D, 0]` indicates
how many padding values to add before the contents of `input` in that dimension,
and `paddings[D, 1]` indicates how many padding values to add after the contents
of `input` in that dimension. `constant_values` is a scalar tensor of the same
type as `input` that indicates the value to use for padding `input`.
The padded size of each dimension D of the output is:
`paddings(D, 0) + input.dim_size(D) + paddings(D, 1)`
For example:
```
# 't' is [[1, 1], [2, 2]]
# 'paddings' is [[1, 1], [2, 2]]
# 'constant_values' is 0
# rank of 't' is 2
pad(t, paddings) ==> [[0, 0, 0, 0, 0, 0]
[0, 0, 1, 1, 0, 0]
[0, 0, 2, 2, 0, 0]
[0, 0, 0, 0, 0, 0]]
```
Args:
input: A `Tensor`.
paddings: A `Tensor`. Must be one of the following types: `int32`, `int64`.
constant_values: A `Tensor`. Must have the same type as `input`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "PadV2", name,
tld.op_callbacks, input, paddings, constant_values)
return _result
except _core._FallbackException:
try:
return pad_v2_eager_fallback(
input, paddings, constant_values, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"PadV2", input=input, paddings=paddings,
constant_values=constant_values, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tpaddings",
_op._get_attr_type("Tpaddings"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"PadV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
PadV2 = tf_export("raw_ops.PadV2")(_ops.to_raw_op(pad_v2))
def pad_v2_eager_fallback(input, paddings, constant_values, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, constant_values], ctx)
(input, constant_values) = _inputs_T
_attr_Tpaddings, (paddings,) = _execute.args_to_matching_eager([paddings], ctx, _dtypes.int32)
_inputs_flat = [input, paddings, constant_values]
_attrs = ("T", _attr_T, "Tpaddings", _attr_Tpaddings)
_result = _execute.execute(b"PadV2", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"PadV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def parallel_concat(values, shape, name=None):
r"""Concatenates a list of `N` tensors along the first dimension.
The input tensors are all required to have size 1 in the first dimension.
For example:
```
# 'x' is [[1, 4]]
# 'y' is [[2, 5]]
# 'z' is [[3, 6]]
parallel_concat([x, y, z]) => [[1, 4], [2, 5], [3, 6]] # Pack along first dim.
```
The difference between concat and parallel_concat is that concat requires all
of the inputs be computed before the operation will begin but doesn't require
that the input shapes be known during graph construction. Parallel concat
will copy pieces of the input into the output as they become available, in
some situations this can provide a performance benefit.
Args:
values: A list of at least 1 `Tensor` objects with the same type.
Tensors to be concatenated. All must have size 1 in the first dimension
and same shape.
shape: A `tf.TensorShape` or list of `ints`.
the final shape of the result; should be equal to the shapes of any input
but with the number of input values in the first dimension.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `values`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ParallelConcat", name,
tld.op_callbacks, values, "shape", shape)
return _result
except _core._FallbackException:
try:
return parallel_concat_eager_fallback(
values, shape=shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'parallel_concat' Op, not %r." % values)
_attr_N = len(values)
shape = _execute.make_shape(shape, "shape")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ParallelConcat", values=values, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"), "T", _op._get_attr_type("T"),
"shape", _op.get_attr("shape"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ParallelConcat", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ParallelConcat = tf_export("raw_ops.ParallelConcat")(_ops.to_raw_op(parallel_concat))
def parallel_concat_eager_fallback(values, shape, name, ctx):
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'parallel_concat' Op, not %r." % values)
_attr_N = len(values)
shape = _execute.make_shape(shape, "shape")
_attr_T, values = _execute.args_to_matching_eager(list(values), ctx)
_inputs_flat = list(values)
_attrs = ("N", _attr_N, "T", _attr_T, "shape", shape)
_result = _execute.execute(b"ParallelConcat", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ParallelConcat", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def placeholder(dtype, shape=None, name=None):
r"""A placeholder op for a value that will be fed into the computation.
N.B. This operation will fail with an error if it is executed. It is
intended as a way to represent a value that will always be fed, and to
provide attrs that enable the fed value to be checked at runtime.
Args:
dtype: A `tf.DType`. The type of elements in the tensor.
shape: An optional `tf.TensorShape` or list of `ints`. Defaults to `None`.
(Optional) The shape of the tensor. If the shape has 0 dimensions, the
shape is unconstrained.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `dtype`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Placeholder", name,
tld.op_callbacks, "dtype", dtype, "shape", shape)
return _result
except _core._FallbackException:
try:
return placeholder_eager_fallback(
dtype=dtype, shape=shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
dtype = _execute.make_type(dtype, "dtype")
if shape is None:
shape = None
shape = _execute.make_shape(shape, "shape")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Placeholder", dtype=dtype, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("dtype", _op._get_attr_type("dtype"), "shape",
_op.get_attr("shape"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Placeholder", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Placeholder = tf_export("raw_ops.Placeholder")(_ops.to_raw_op(placeholder))
def placeholder_eager_fallback(dtype, shape, name, ctx):
dtype = _execute.make_type(dtype, "dtype")
if shape is None:
shape = None
shape = _execute.make_shape(shape, "shape")
_inputs_flat = []
_attrs = ("dtype", dtype, "shape", shape)
_result = _execute.execute(b"Placeholder", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Placeholder", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def placeholder_v2(dtype, shape, name=None):
r"""A placeholder op for a value that will be fed into the computation.
N.B. This operation will fail with an error if it is executed. It is
intended as a way to represent a value that will always be fed, and to
provide attrs that enable the fed value to be checked at runtime.
Args:
dtype: A `tf.DType`. The type of elements in the tensor.
shape: A `tf.TensorShape` or list of `ints`.
The shape of the tensor. The shape can be any partially-specified
shape. To be unconstrained, pass in a shape with unknown rank.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `dtype`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "PlaceholderV2", name,
tld.op_callbacks, "dtype", dtype, "shape", shape)
return _result
except _core._FallbackException:
try:
return placeholder_v2_eager_fallback(
dtype=dtype, shape=shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
dtype = _execute.make_type(dtype, "dtype")
shape = _execute.make_shape(shape, "shape")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"PlaceholderV2", dtype=dtype, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("dtype", _op._get_attr_type("dtype"), "shape",
_op.get_attr("shape"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"PlaceholderV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
PlaceholderV2 = tf_export("raw_ops.PlaceholderV2")(_ops.to_raw_op(placeholder_v2))
def placeholder_v2_eager_fallback(dtype, shape, name, ctx):
dtype = _execute.make_type(dtype, "dtype")
shape = _execute.make_shape(shape, "shape")
_inputs_flat = []
_attrs = ("dtype", dtype, "shape", shape)
_result = _execute.execute(b"PlaceholderV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"PlaceholderV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def placeholder_with_default(input, shape, name=None):
r"""A placeholder op that passes through `input` when its output is not fed.
Args:
input: A `Tensor`. The default value to produce when `output` is not fed.
shape: A `tf.TensorShape` or list of `ints`.
The (possibly partial) shape of the tensor.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "PlaceholderWithDefault", name,
tld.op_callbacks, input, "shape", shape)
return _result
except _core._FallbackException:
try:
return placeholder_with_default_eager_fallback(
input, shape=shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
shape = _execute.make_shape(shape, "shape")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"PlaceholderWithDefault", input=input, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("dtype", _op._get_attr_type("dtype"), "shape",
_op.get_attr("shape"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"PlaceholderWithDefault", _inputs_flat, _attrs, _result)
_result, = _result
return _result
PlaceholderWithDefault = tf_export("raw_ops.PlaceholderWithDefault")(_ops.to_raw_op(placeholder_with_default))
def placeholder_with_default_eager_fallback(input, shape, name, ctx):
shape = _execute.make_shape(shape, "shape")
_attr_dtype, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("dtype", _attr_dtype, "shape", shape)
_result = _execute.execute(b"PlaceholderWithDefault", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"PlaceholderWithDefault", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def prevent_gradient(input, message="", name=None):
r"""An identity op that triggers an error if a gradient is requested.
When executed in a graph, this op outputs its input tensor as-is.
When building ops to compute gradients, the TensorFlow gradient system
will return an error when trying to lookup the gradient of this op,
because no gradient must ever be registered for this function. This
op exists to prevent subtle bugs from silently returning unimplemented
gradients in some corner cases.
Args:
input: A `Tensor`. any tensor.
message: An optional `string`. Defaults to `""`.
Will be printed in the error when anyone tries to differentiate
this operation.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "PreventGradient", name,
tld.op_callbacks, input, "message", message)
return _result
except _core._FallbackException:
try:
return prevent_gradient_eager_fallback(
input, message=message, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if message is None:
message = ""
message = _execute.make_str(message, "message")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"PreventGradient", input=input, message=message, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "message",
_op.get_attr("message"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"PreventGradient", _inputs_flat, _attrs, _result)
_result, = _result
return _result
PreventGradient = tf_export("raw_ops.PreventGradient")(_ops.to_raw_op(prevent_gradient))
def prevent_gradient_eager_fallback(input, message, name, ctx):
if message is None:
message = ""
message = _execute.make_str(message, "message")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "message", message)
_result = _execute.execute(b"PreventGradient", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"PreventGradient", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def quantize_and_dequantize(input, signed_input=True, num_bits=8, range_given=False, input_min=0, input_max=0, name=None):
r"""Use QuantizeAndDequantizeV2 instead.
Args:
input: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`.
signed_input: An optional `bool`. Defaults to `True`.
num_bits: An optional `int`. Defaults to `8`.
range_given: An optional `bool`. Defaults to `False`.
input_min: An optional `float`. Defaults to `0`.
input_max: An optional `float`. Defaults to `0`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizeAndDequantize", name,
tld.op_callbacks, input, "signed_input", signed_input, "num_bits",
num_bits, "range_given", range_given, "input_min", input_min,
"input_max", input_max)
return _result
except _core._FallbackException:
try:
return quantize_and_dequantize_eager_fallback(
input, signed_input=signed_input, num_bits=num_bits,
range_given=range_given, input_min=input_min, input_max=input_max,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if signed_input is None:
signed_input = True
signed_input = _execute.make_bool(signed_input, "signed_input")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if range_given is None:
range_given = False
range_given = _execute.make_bool(range_given, "range_given")
if input_min is None:
input_min = 0
input_min = _execute.make_float(input_min, "input_min")
if input_max is None:
input_max = 0
input_max = _execute.make_float(input_max, "input_max")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizeAndDequantize", input=input, signed_input=signed_input,
num_bits=num_bits, range_given=range_given,
input_min=input_min, input_max=input_max,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("signed_input", _op._get_attr_bool("signed_input"), "num_bits",
_op._get_attr_int("num_bits"), "range_given",
_op._get_attr_bool("range_given"), "input_min",
_op.get_attr("input_min"), "input_max",
_op.get_attr("input_max"), "T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizeAndDequantize", _inputs_flat, _attrs, _result)
_result, = _result
return _result
QuantizeAndDequantize = tf_export("raw_ops.QuantizeAndDequantize")(_ops.to_raw_op(quantize_and_dequantize))
def quantize_and_dequantize_eager_fallback(input, signed_input, num_bits, range_given, input_min, input_max, name, ctx):
if signed_input is None:
signed_input = True
signed_input = _execute.make_bool(signed_input, "signed_input")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if range_given is None:
range_given = False
range_given = _execute.make_bool(range_given, "range_given")
if input_min is None:
input_min = 0
input_min = _execute.make_float(input_min, "input_min")
if input_max is None:
input_max = 0
input_max = _execute.make_float(input_max, "input_max")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("signed_input", signed_input, "num_bits", num_bits, "range_given",
range_given, "input_min", input_min, "input_max", input_max, "T", _attr_T)
_result = _execute.execute(b"QuantizeAndDequantize", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizeAndDequantize", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def quantize_and_dequantize_v2(input, input_min, input_max, signed_input=True, num_bits=8, range_given=False, round_mode="HALF_TO_EVEN", narrow_range=False, axis=-1, name=None):
r"""Quantizes then dequantizes a tensor.
This op simulates the precision loss from the quantized forward pass by:
1. Quantizing the tensor to fixed point numbers, which should match the target
quantization method when it is used in inference.
2. Dequantizing it back to floating point numbers for the following ops, most
likely matmul.
There are different ways to quantize. This version uses only scaling, so 0.0
maps to 0.
From the specified 'num_bits' in the quantized output type, it determines
minimum and maximum representable quantized values.
e.g.
* [-128, 127] for signed, num_bits = 8, or
* [0, 255] for unsigned, num_bits = 8.
If range_given == False, the initial input_min, input_max will be determined
automatically as the minimum and maximum values in the input tensor, otherwise
the specified values of input_min, input_max are used.
Note: If the input_min, input_max are specified, they do not need to equal the
actual minimum and maximum values in the tensor. e.g. in some cases it may be
beneficial to specify these values such that the low probability extremes of the
input distribution are clipped.
This op determines the maximum scale_factor that would map the initial
[input_min, input_max] range to a range that lies within the representable
quantized range.
It determines the scale from one of input_min and input_max, then updates the
other one to maximize the representable range.
e.g.
* if the output is signed, num_bits = 8, [input_min, input_max] = [-10.0,
5.0]: it would use a scale_factor of -128 / -10.0 = 12.8 In this case, it
would update input_max to be 127 / 12.8 = 9.921875
* if the output is signed, num_bits = 8, [input_min, input_max] = [-10.0,
10.0]: it would use a scale_factor of 127 / 10.0 = 12.7 In this case, it
would update input_min to be 128.0 / 12.7 = -10.07874
* if the output is unsigned, input_min is forced to be 0, and only the
specified input_max is used.
After determining the scale_factor and updating the input range, it applies the
following to each value in the 'input' tensor.
output = round(clamp(value, input_min, input_max) * scale_factor) / scale_factor.
The above round function rounds the value based on the given round_mode.
Args:
input: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`.
Tensor to quantize and then dequantize.
input_min: A `Tensor`. Must have the same type as `input`.
If `range_given == True`, this specifies the minimum input value that needs to
be represented, otherwise it is determined from the min value of the `input`
tensor.
input_max: A `Tensor`. Must have the same type as `input`.
If `range_given == True`, this specifies the maximum input value that needs to
be represented, otherwise it is determined from the max value of the `input`
tensor.
signed_input: An optional `bool`. Defaults to `True`.
Whether the quantization is signed or unsigned. (actually this parameter should
have been called <b>`signed_output`</b>)
num_bits: An optional `int`. Defaults to `8`.
The bitwidth of the quantization.
range_given: An optional `bool`. Defaults to `False`.
Whether the range is given or should be determined from the `input` tensor.
round_mode: An optional `string` from: `"HALF_TO_EVEN", "HALF_UP"`. Defaults to `"HALF_TO_EVEN"`.
The 'round_mode' attribute controls which rounding tie-breaking algorithm is
used when rounding float values to their quantized equivalents. The following
rounding modes are currently supported:
* HALF_TO_EVEN: this is the default round_mode.
* HALF_UP: round towards positive. In this mode 7.5 rounds up to 8 and -7.5
rounds up to -7.
narrow_range: An optional `bool`. Defaults to `False`.
If True, then the absolute value of the quantized minimum value is the same as
the quantized maximum value, instead of 1 greater.
i.e. for 8 bit quantization, the minimum value is -127 instead of -128.
axis: An optional `int`. Defaults to `-1`.
If specified, this axis is treated as a channel or slice axis, and a separate
quantization range is used for each channel or slice along this axis.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizeAndDequantizeV2",
name, tld.op_callbacks, input, input_min, input_max, "signed_input",
signed_input, "num_bits", num_bits, "range_given", range_given,
"round_mode", round_mode, "narrow_range", narrow_range, "axis", axis)
return _result
except _core._FallbackException:
try:
return quantize_and_dequantize_v2_eager_fallback(
input, input_min, input_max, signed_input=signed_input,
num_bits=num_bits, range_given=range_given, round_mode=round_mode,
narrow_range=narrow_range, axis=axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if signed_input is None:
signed_input = True
signed_input = _execute.make_bool(signed_input, "signed_input")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if range_given is None:
range_given = False
range_given = _execute.make_bool(range_given, "range_given")
if round_mode is None:
round_mode = "HALF_TO_EVEN"
round_mode = _execute.make_str(round_mode, "round_mode")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizeAndDequantizeV2", input=input, input_min=input_min,
input_max=input_max,
signed_input=signed_input,
num_bits=num_bits, range_given=range_given,
round_mode=round_mode,
narrow_range=narrow_range, axis=axis,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("signed_input", _op._get_attr_bool("signed_input"), "num_bits",
_op._get_attr_int("num_bits"), "range_given",
_op._get_attr_bool("range_given"), "T", _op._get_attr_type("T"),
"round_mode", _op.get_attr("round_mode"), "narrow_range",
_op._get_attr_bool("narrow_range"), "axis",
_op._get_attr_int("axis"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizeAndDequantizeV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
QuantizeAndDequantizeV2 = tf_export("raw_ops.QuantizeAndDequantizeV2")(_ops.to_raw_op(quantize_and_dequantize_v2))
def quantize_and_dequantize_v2_eager_fallback(input, input_min, input_max, signed_input, num_bits, range_given, round_mode, narrow_range, axis, name, ctx):
if signed_input is None:
signed_input = True
signed_input = _execute.make_bool(signed_input, "signed_input")
if num_bits is None:
num_bits = 8
num_bits = _execute.make_int(num_bits, "num_bits")
if range_given is None:
range_given = False
range_given = _execute.make_bool(range_given, "range_given")
if round_mode is None:
round_mode = "HALF_TO_EVEN"
round_mode = _execute.make_str(round_mode, "round_mode")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, input_min, input_max], ctx)
(input, input_min, input_max) = _inputs_T
_inputs_flat = [input, input_min, input_max]
_attrs = ("signed_input", signed_input, "num_bits", num_bits, "range_given",
range_given, "T", _attr_T, "round_mode", round_mode, "narrow_range",
narrow_range, "axis", axis)
_result = _execute.execute(b"QuantizeAndDequantizeV2", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizeAndDequantizeV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def quantize_and_dequantize_v3(input, input_min, input_max, num_bits, signed_input=True, range_given=True, narrow_range=False, axis=-1, name=None):
r"""Quantizes then dequantizes a tensor.
This is almost identical to QuantizeAndDequantizeV2, except that num_bits is a
tensor, so its value can change during training.
Args:
input: A `Tensor`. Must be one of the following types: `bfloat16`, `half`, `float32`, `float64`.
input_min: A `Tensor`. Must have the same type as `input`.
input_max: A `Tensor`. Must have the same type as `input`.
num_bits: A `Tensor` of type `int32`.
signed_input: An optional `bool`. Defaults to `True`.
range_given: An optional `bool`. Defaults to `True`.
narrow_range: An optional `bool`. Defaults to `False`.
axis: An optional `int`. Defaults to `-1`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizeAndDequantizeV3",
name, tld.op_callbacks, input, input_min, input_max, num_bits,
"signed_input", signed_input, "range_given", range_given,
"narrow_range", narrow_range, "axis", axis)
return _result
except _core._FallbackException:
try:
return quantize_and_dequantize_v3_eager_fallback(
input, input_min, input_max, num_bits, signed_input=signed_input,
range_given=range_given, narrow_range=narrow_range, axis=axis,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if signed_input is None:
signed_input = True
signed_input = _execute.make_bool(signed_input, "signed_input")
if range_given is None:
range_given = True
range_given = _execute.make_bool(range_given, "range_given")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizeAndDequantizeV3", input=input, input_min=input_min,
input_max=input_max, num_bits=num_bits,
signed_input=signed_input,
range_given=range_given,
narrow_range=narrow_range, axis=axis,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("signed_input", _op._get_attr_bool("signed_input"),
"range_given", _op._get_attr_bool("range_given"), "T",
_op._get_attr_type("T"), "narrow_range",
_op._get_attr_bool("narrow_range"), "axis",
_op._get_attr_int("axis"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizeAndDequantizeV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
QuantizeAndDequantizeV3 = tf_export("raw_ops.QuantizeAndDequantizeV3")(_ops.to_raw_op(quantize_and_dequantize_v3))
def quantize_and_dequantize_v3_eager_fallback(input, input_min, input_max, num_bits, signed_input, range_given, narrow_range, axis, name, ctx):
if signed_input is None:
signed_input = True
signed_input = _execute.make_bool(signed_input, "signed_input")
if range_given is None:
range_given = True
range_given = _execute.make_bool(range_given, "range_given")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, input_min, input_max], ctx)
(input, input_min, input_max) = _inputs_T
num_bits = _ops.convert_to_tensor(num_bits, _dtypes.int32)
_inputs_flat = [input, input_min, input_max, num_bits]
_attrs = ("signed_input", signed_input, "range_given", range_given, "T",
_attr_T, "narrow_range", narrow_range, "axis", axis)
_result = _execute.execute(b"QuantizeAndDequantizeV3", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizeAndDequantizeV3", _inputs_flat, _attrs, _result)
_result, = _result
return _result
_QuantizeV2Output = collections.namedtuple(
"QuantizeV2",
["output", "output_min", "output_max"])
def quantize_v2(input, min_range, max_range, T, mode="MIN_COMBINED", round_mode="HALF_AWAY_FROM_ZERO", narrow_range=False, axis=-1, ensure_minimum_range=0.01, name=None):
r"""Quantize the 'input' tensor of type float to 'output' tensor of type 'T'.
[min_range, max_range] are scalar floats that specify the range for
the 'input' data. The 'mode' attribute controls exactly which calculations are
used to convert the float values to their quantized equivalents. The
'round_mode' attribute controls which rounding tie-breaking algorithm is used
when rounding float values to their quantized equivalents.
In 'MIN_COMBINED' mode, each value of the tensor will undergo the following:
```
out[i] = (in[i] - min_range) * range(T) / (max_range - min_range)
if T == qint8: out[i] -= (range(T) + 1) / 2.0
```
here `range(T) = numeric_limits<T>::max() - numeric_limits<T>::min()`
*MIN_COMBINED Mode Example*
Assume the input is type float and has a possible range of [0.0, 6.0] and the
output type is quint8 ([0, 255]). The min_range and max_range values should be
specified as 0.0 and 6.0. Quantizing from float to quint8 will multiply each
value of the input by 255/6 and cast to quint8.
If the output type was qint8 ([-128, 127]), the operation will additionally
subtract each value by 128 prior to casting, so that the range of values aligns
with the range of qint8.
If the mode is 'MIN_FIRST', then this approach is used:
```
num_discrete_values = 1 << (# of bits in T)
range_adjust = num_discrete_values / (num_discrete_values - 1)
range = (range_max - range_min) * range_adjust
range_scale = num_discrete_values / range
quantized = round(input * range_scale) - round(range_min * range_scale) +
numeric_limits<T>::min()
quantized = max(quantized, numeric_limits<T>::min())
quantized = min(quantized, numeric_limits<T>::max())
```
The biggest difference between this and MIN_COMBINED is that the minimum range
is rounded first, before it's subtracted from the rounded value. With
MIN_COMBINED, a small bias is introduced where repeated iterations of quantizing
and dequantizing will introduce a larger and larger error.
*SCALED mode Example*
`SCALED` mode matches the quantization approach used in
`QuantizeAndDequantize{V2|V3}`.
If the mode is `SCALED`, the quantization is performed by multiplying each
input value by a scaling_factor.
The scaling_factor is determined from `min_range` and `max_range` to be as large
as possible such that the range from `min_range` to `max_range` is representable
within values of type T.
```c++
const int min_T = std::numeric_limits<T>::min();
const int max_T = std::numeric_limits<T>::max();
const float max_float = std::numeric_limits<float>::max();
const float scale_factor_from_min_side =
(min_T * min_range > 0) ? min_T / min_range : max_float;
const float scale_factor_from_max_side =
(max_T * max_range > 0) ? max_T / max_range : max_float;
const float scale_factor = std::min(scale_factor_from_min_side,
scale_factor_from_max_side);
```
We next use the scale_factor to adjust min_range and max_range as follows:
```c++
min_range = min_T / scale_factor;
max_range = max_T / scale_factor;
```
e.g. if T = qint8, and initially min_range = -10, and max_range = 9, we would
compare -128/-10.0 = 12.8 to 127/9.0 = 14.11, and set scaling_factor = 12.8
In this case, min_range would remain -10, but max_range would be adjusted to
127 / 12.8 = 9.921875
So we will quantize input values in the range (-10, 9.921875) to (-128, 127).
The input tensor can now be quantized by clipping values to the range
`min_range` to `max_range`, then multiplying by scale_factor as follows:
```c++
result = round(min(max_range, max(min_range, input)) * scale_factor)
```
The adjusted `min_range` and `max_range` are returned as outputs 2 and 3 of
this operation. These outputs should be used as the range for any further
calculations.
*narrow_range (bool) attribute*
If true, we do not use the minimum quantized value.
i.e. for int8 the quantized output, it would be restricted to the range
-127..127 instead of the full -128..127 range.
This is provided for compatibility with certain inference backends.
(Only applies to SCALED mode)
*axis (int) attribute*
An optional `axis` attribute can specify a dimension index of the input tensor,
such that quantization ranges will be calculated and applied separately for each
slice of the tensor along that dimension. This is useful for per-channel
quantization.
If axis is specified, min_range and max_range
if `axis`=None, per-tensor quantization is performed as normal.
*ensure_minimum_range (float) attribute*
Ensures the minimum quantization range is at least this value.
The legacy default value for this is 0.01, but it is strongly suggested to
set it to 0 for new uses.
Args:
input: A `Tensor` of type `float32`.
min_range: A `Tensor` of type `float32`.
The minimum value of the quantization range. This value may be adjusted by the
op depending on other parameters. The adjusted value is written to `output_min`.
If the `axis` attribute is specified, this must be a 1-D tensor whose size
matches the `axis` dimension of the input and output tensors.
max_range: A `Tensor` of type `float32`.
The maximum value of the quantization range. This value may be adjusted by the
op depending on other parameters. The adjusted value is written to `output_max`.
If the `axis` attribute is specified, this must be a 1-D tensor whose size
matches the `axis` dimension of the input and output tensors.
T: A `tf.DType` from: `tf.qint8, tf.quint8, tf.qint32, tf.qint16, tf.quint16`.
mode: An optional `string` from: `"MIN_COMBINED", "MIN_FIRST", "SCALED"`. Defaults to `"MIN_COMBINED"`.
round_mode: An optional `string` from: `"HALF_AWAY_FROM_ZERO", "HALF_TO_EVEN"`. Defaults to `"HALF_AWAY_FROM_ZERO"`.
narrow_range: An optional `bool`. Defaults to `False`.
axis: An optional `int`. Defaults to `-1`.
ensure_minimum_range: An optional `float`. Defaults to `0.01`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (output, output_min, output_max).
output: A `Tensor` of type `T`.
output_min: A `Tensor` of type `float32`.
output_max: A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizeV2", name,
tld.op_callbacks, input, min_range, max_range, "T", T, "mode", mode,
"round_mode", round_mode, "narrow_range", narrow_range, "axis", axis,
"ensure_minimum_range", ensure_minimum_range)
_result = _QuantizeV2Output._make(_result)
return _result
except _core._FallbackException:
try:
return quantize_v2_eager_fallback(
input, min_range, max_range, T=T, mode=mode,
round_mode=round_mode, narrow_range=narrow_range, axis=axis,
ensure_minimum_range=ensure_minimum_range, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
T = _execute.make_type(T, "T")
if mode is None:
mode = "MIN_COMBINED"
mode = _execute.make_str(mode, "mode")
if round_mode is None:
round_mode = "HALF_AWAY_FROM_ZERO"
round_mode = _execute.make_str(round_mode, "round_mode")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
if ensure_minimum_range is None:
ensure_minimum_range = 0.01
ensure_minimum_range = _execute.make_float(ensure_minimum_range, "ensure_minimum_range")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizeV2", input=input, min_range=min_range, max_range=max_range,
T=T, mode=mode, round_mode=round_mode,
narrow_range=narrow_range, axis=axis,
ensure_minimum_range=ensure_minimum_range, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "mode", _op.get_attr("mode"),
"round_mode", _op.get_attr("round_mode"), "narrow_range",
_op._get_attr_bool("narrow_range"), "axis",
_op._get_attr_int("axis"), "ensure_minimum_range",
_op.get_attr("ensure_minimum_range"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizeV2", _inputs_flat, _attrs, _result)
_result = _QuantizeV2Output._make(_result)
return _result
QuantizeV2 = tf_export("raw_ops.QuantizeV2")(_ops.to_raw_op(quantize_v2))
def quantize_v2_eager_fallback(input, min_range, max_range, T, mode, round_mode, narrow_range, axis, ensure_minimum_range, name, ctx):
T = _execute.make_type(T, "T")
if mode is None:
mode = "MIN_COMBINED"
mode = _execute.make_str(mode, "mode")
if round_mode is None:
round_mode = "HALF_AWAY_FROM_ZERO"
round_mode = _execute.make_str(round_mode, "round_mode")
if narrow_range is None:
narrow_range = False
narrow_range = _execute.make_bool(narrow_range, "narrow_range")
if axis is None:
axis = -1
axis = _execute.make_int(axis, "axis")
if ensure_minimum_range is None:
ensure_minimum_range = 0.01
ensure_minimum_range = _execute.make_float(ensure_minimum_range, "ensure_minimum_range")
input = _ops.convert_to_tensor(input, _dtypes.float32)
min_range = _ops.convert_to_tensor(min_range, _dtypes.float32)
max_range = _ops.convert_to_tensor(max_range, _dtypes.float32)
_inputs_flat = [input, min_range, max_range]
_attrs = ("T", T, "mode", mode, "round_mode", round_mode, "narrow_range",
narrow_range, "axis", axis, "ensure_minimum_range", ensure_minimum_range)
_result = _execute.execute(b"QuantizeV2", 3, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizeV2", _inputs_flat, _attrs, _result)
_result = _QuantizeV2Output._make(_result)
return _result
_QuantizedConcatOutput = collections.namedtuple(
"QuantizedConcat",
["output", "output_min", "output_max"])
@_dispatch.add_dispatch_list
@tf_export('quantization.quantized_concat', v1=['quantization.quantized_concat', 'quantized_concat'])
@deprecated_endpoints('quantized_concat')
def quantized_concat(concat_dim, values, input_mins, input_maxes, name=None):
r"""Concatenates quantized tensors along one dimension.
Args:
concat_dim: A `Tensor` of type `int32`.
0-D. The dimension along which to concatenate. Must be in the
range [0, rank(values)).
values: A list of at least 2 `Tensor` objects with the same type.
The `N` Tensors to concatenate. Their ranks and types must match,
and their sizes must match in all dimensions except `concat_dim`.
input_mins: A list with the same length as `values` of `Tensor` objects with type `float32`.
The minimum scalar values for each of the input tensors.
input_maxes: A list with the same length as `values` of `Tensor` objects with type `float32`.
The maximum scalar values for each of the input tensors.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (output, output_min, output_max).
output: A `Tensor`. Has the same type as `values`.
output_min: A `Tensor` of type `float32`.
output_max: A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizedConcat", name,
tld.op_callbacks, concat_dim, values, input_mins, input_maxes)
_result = _QuantizedConcatOutput._make(_result)
return _result
except _core._FallbackException:
try:
return quantized_concat_eager_fallback(
concat_dim, values, input_mins, input_maxes, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
quantized_concat, concat_dim=concat_dim, values=values,
input_mins=input_mins,
input_maxes=input_maxes, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'quantized_concat' Op, not %r." % values)
_attr_N = len(values)
if not isinstance(input_mins, (list, tuple)):
raise TypeError(
"Expected list for 'input_mins' argument to "
"'quantized_concat' Op, not %r." % input_mins)
if len(input_mins) != _attr_N:
raise ValueError(
"List argument 'input_mins' to 'quantized_concat' Op with length %d "
"must match length %d of argument 'values'." %
(len(input_mins), _attr_N))
if not isinstance(input_maxes, (list, tuple)):
raise TypeError(
"Expected list for 'input_maxes' argument to "
"'quantized_concat' Op, not %r." % input_maxes)
if len(input_maxes) != _attr_N:
raise ValueError(
"List argument 'input_maxes' to 'quantized_concat' Op with length %d "
"must match length %d of argument 'values'." %
(len(input_maxes), _attr_N))
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizedConcat", concat_dim=concat_dim, values=values,
input_mins=input_mins, input_maxes=input_maxes,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
quantized_concat, concat_dim=concat_dim, values=values,
input_mins=input_mins, input_maxes=input_maxes,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"), "T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizedConcat", _inputs_flat, _attrs, _result)
_result = _QuantizedConcatOutput._make(_result)
return _result
QuantizedConcat = tf_export("raw_ops.QuantizedConcat")(_ops.to_raw_op(quantized_concat))
def quantized_concat_eager_fallback(concat_dim, values, input_mins, input_maxes, name, ctx):
if not isinstance(values, (list, tuple)):
raise TypeError(
"Expected list for 'values' argument to "
"'quantized_concat' Op, not %r." % values)
_attr_N = len(values)
if not isinstance(input_mins, (list, tuple)):
raise TypeError(
"Expected list for 'input_mins' argument to "
"'quantized_concat' Op, not %r." % input_mins)
if len(input_mins) != _attr_N:
raise ValueError(
"List argument 'input_mins' to 'quantized_concat' Op with length %d "
"must match length %d of argument 'values'." %
(len(input_mins), _attr_N))
if not isinstance(input_maxes, (list, tuple)):
raise TypeError(
"Expected list for 'input_maxes' argument to "
"'quantized_concat' Op, not %r." % input_maxes)
if len(input_maxes) != _attr_N:
raise ValueError(
"List argument 'input_maxes' to 'quantized_concat' Op with length %d "
"must match length %d of argument 'values'." %
(len(input_maxes), _attr_N))
_attr_T, values = _execute.args_to_matching_eager(list(values), ctx)
concat_dim = _ops.convert_to_tensor(concat_dim, _dtypes.int32)
input_mins = _ops.convert_n_to_tensor(input_mins, _dtypes.float32)
input_maxes = _ops.convert_n_to_tensor(input_maxes, _dtypes.float32)
_inputs_flat = [concat_dim] + list(values) + list(input_mins) + list(input_maxes)
_attrs = ("N", _attr_N, "T", _attr_T)
_result = _execute.execute(b"QuantizedConcat", 3, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizedConcat", _inputs_flat, _attrs, _result)
_result = _QuantizedConcatOutput._make(_result)
return _result
_QuantizedInstanceNormOutput = collections.namedtuple(
"QuantizedInstanceNorm",
["y", "y_min", "y_max"])
def quantized_instance_norm(x, x_min, x_max, output_range_given=False, given_y_min=0, given_y_max=0, variance_epsilon=1e-05, min_separation=0.001, name=None):
r"""Quantized Instance normalization.
Args:
x: A `Tensor`. Must be one of the following types: `qint8`, `quint8`, `qint32`, `qint16`, `quint16`.
A 4D input Tensor.
x_min: A `Tensor` of type `float32`.
The value represented by the lowest quantized input.
x_max: A `Tensor` of type `float32`.
The value represented by the highest quantized input.
output_range_given: An optional `bool`. Defaults to `False`.
If True, `given_y_min` and `given_y_min`
and `given_y_max` are used as the output range. Otherwise,
the implementation computes the output range.
given_y_min: An optional `float`. Defaults to `0`.
Output in `y_min` if `output_range_given` is True.
given_y_max: An optional `float`. Defaults to `0`.
Output in `y_max` if `output_range_given` is True.
variance_epsilon: An optional `float`. Defaults to `1e-05`.
A small float number to avoid dividing by 0.
min_separation: An optional `float`. Defaults to `0.001`.
Minimum value of `y_max - y_min`
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (y, y_min, y_max).
y: A `Tensor`. Has the same type as `x`.
y_min: A `Tensor` of type `float32`.
y_max: A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizedInstanceNorm", name,
tld.op_callbacks, x, x_min, x_max, "output_range_given",
output_range_given, "given_y_min", given_y_min, "given_y_max",
given_y_max, "variance_epsilon", variance_epsilon, "min_separation",
min_separation)
_result = _QuantizedInstanceNormOutput._make(_result)
return _result
except _core._FallbackException:
try:
return quantized_instance_norm_eager_fallback(
x, x_min, x_max, output_range_given=output_range_given,
given_y_min=given_y_min, given_y_max=given_y_max,
variance_epsilon=variance_epsilon, min_separation=min_separation,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if output_range_given is None:
output_range_given = False
output_range_given = _execute.make_bool(output_range_given, "output_range_given")
if given_y_min is None:
given_y_min = 0
given_y_min = _execute.make_float(given_y_min, "given_y_min")
if given_y_max is None:
given_y_max = 0
given_y_max = _execute.make_float(given_y_max, "given_y_max")
if variance_epsilon is None:
variance_epsilon = 1e-05
variance_epsilon = _execute.make_float(variance_epsilon, "variance_epsilon")
if min_separation is None:
min_separation = 0.001
min_separation = _execute.make_float(min_separation, "min_separation")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizedInstanceNorm", x=x, x_min=x_min, x_max=x_max,
output_range_given=output_range_given,
given_y_min=given_y_min,
given_y_max=given_y_max,
variance_epsilon=variance_epsilon,
min_separation=min_separation, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "output_range_given",
_op._get_attr_bool("output_range_given"), "given_y_min",
_op.get_attr("given_y_min"), "given_y_max",
_op.get_attr("given_y_max"), "variance_epsilon",
_op.get_attr("variance_epsilon"), "min_separation",
_op.get_attr("min_separation"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizedInstanceNorm", _inputs_flat, _attrs, _result)
_result = _QuantizedInstanceNormOutput._make(_result)
return _result
QuantizedInstanceNorm = tf_export("raw_ops.QuantizedInstanceNorm")(_ops.to_raw_op(quantized_instance_norm))
def quantized_instance_norm_eager_fallback(x, x_min, x_max, output_range_given, given_y_min, given_y_max, variance_epsilon, min_separation, name, ctx):
if output_range_given is None:
output_range_given = False
output_range_given = _execute.make_bool(output_range_given, "output_range_given")
if given_y_min is None:
given_y_min = 0
given_y_min = _execute.make_float(given_y_min, "given_y_min")
if given_y_max is None:
given_y_max = 0
given_y_max = _execute.make_float(given_y_max, "given_y_max")
if variance_epsilon is None:
variance_epsilon = 1e-05
variance_epsilon = _execute.make_float(variance_epsilon, "variance_epsilon")
if min_separation is None:
min_separation = 0.001
min_separation = _execute.make_float(min_separation, "min_separation")
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
x_min = _ops.convert_to_tensor(x_min, _dtypes.float32)
x_max = _ops.convert_to_tensor(x_max, _dtypes.float32)
_inputs_flat = [x, x_min, x_max]
_attrs = ("T", _attr_T, "output_range_given", output_range_given,
"given_y_min", given_y_min, "given_y_max", given_y_max, "variance_epsilon",
variance_epsilon, "min_separation", min_separation)
_result = _execute.execute(b"QuantizedInstanceNorm", 3, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizedInstanceNorm", _inputs_flat, _attrs, _result)
_result = _QuantizedInstanceNormOutput._make(_result)
return _result
_QuantizedReshapeOutput = collections.namedtuple(
"QuantizedReshape",
["output", "output_min", "output_max"])
def quantized_reshape(tensor, shape, input_min, input_max, name=None):
r"""Reshapes a quantized tensor as per the Reshape op.
```
Args:
tensor: A `Tensor`.
shape: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Defines the shape of the output tensor.
input_min: A `Tensor` of type `float32`. The minimum value of the input.
input_max: A `Tensor` of type `float32`. The maximum value of the input.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (output, output_min, output_max).
output: A `Tensor`. Has the same type as `tensor`.
output_min: A `Tensor` of type `float32`.
output_max: A `Tensor` of type `float32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "QuantizedReshape", name,
tld.op_callbacks, tensor, shape, input_min, input_max)
_result = _QuantizedReshapeOutput._make(_result)
return _result
except _core._FallbackException:
try:
return quantized_reshape_eager_fallback(
tensor, shape, input_min, input_max, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"QuantizedReshape", tensor=tensor, shape=shape, input_min=input_min,
input_max=input_max, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tshape",
_op._get_attr_type("Tshape"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"QuantizedReshape", _inputs_flat, _attrs, _result)
_result = _QuantizedReshapeOutput._make(_result)
return _result
QuantizedReshape = tf_export("raw_ops.QuantizedReshape")(_ops.to_raw_op(quantized_reshape))
def quantized_reshape_eager_fallback(tensor, shape, input_min, input_max, name, ctx):
_attr_T, (tensor,) = _execute.args_to_matching_eager([tensor], ctx)
_attr_Tshape, (shape,) = _execute.args_to_matching_eager([shape], ctx, _dtypes.int32)
input_min = _ops.convert_to_tensor(input_min, _dtypes.float32)
input_max = _ops.convert_to_tensor(input_max, _dtypes.float32)
_inputs_flat = [tensor, shape, input_min, input_max]
_attrs = ("T", _attr_T, "Tshape", _attr_Tshape)
_result = _execute.execute(b"QuantizedReshape", 3, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"QuantizedReshape", _inputs_flat, _attrs, _result)
_result = _QuantizedReshapeOutput._make(_result)
return _result
def rank(input, name=None):
r"""Returns the rank of a tensor.
This operation returns an integer representing the rank of `input`.
For example:
```
# 't' is [[[1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]]]
# shape of tensor 't' is [2, 2, 3]
rank(t) ==> 3
```
**Note**: The rank of a tensor is not the same as the rank of a matrix. The rank
of a tensor is the number of indices required to uniquely select each element
of the tensor. Rank is also known as "order", "degree", or "ndims."
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `int32`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Rank", name, tld.op_callbacks,
input)
return _result
except _core._FallbackException:
try:
return rank_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Rank", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Rank", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Rank = tf_export("raw_ops.Rank")(_ops.to_raw_op(rank))
def rank_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Rank", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Rank", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def ref_identity(input, name=None):
r"""Return the same ref tensor as the input ref tensor.
Args:
input: A mutable `Tensor`.
name: A name for the operation (optional).
Returns:
A mutable `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
raise RuntimeError("ref_identity op does not support eager execution. Arg 'output' is a ref.")
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"RefIdentity", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"RefIdentity", _inputs_flat, _attrs, _result)
_result, = _result
return _result
RefIdentity = tf_export("raw_ops.RefIdentity")(_ops.to_raw_op(ref_identity))
def ref_identity_eager_fallback(input, name, ctx):
raise RuntimeError("ref_identity op does not support eager execution. Arg 'output' is a ref.")
def reshape(tensor, shape, name=None):
r"""Reshapes a tensor.
Given `tensor`, this operation returns a tensor that has the same values
as `tensor` with shape `shape`.
If one component of 1-D tensor `shape` is the special value -1, the size of that
dimension is computed so that the total size remains constant. In particular, a
`shape` of `[-1]` flattens into 1-D. At most one component of `shape` may be
unknown.
The `shape` must be 1-D and the operation returns a tensor with shape
`shape` filled with the values of `tensor`. In this case, the number of elements
implied by `shape` must be the same as the number of elements in `tensor`.
It is an error if `shape` is not 1-D.
For example:
```
# tensor 't' is [1, 2, 3, 4, 5, 6, 7, 8, 9]
# tensor 't' has shape [9]
reshape(t, [3, 3]) ==> [[1, 2, 3],
[4, 5, 6],
[7, 8, 9]]
# tensor 't' is [[[1, 1], [2, 2]],
# [[3, 3], [4, 4]]]
# tensor 't' has shape [2, 2, 2]
reshape(t, [2, 4]) ==> [[1, 1, 2, 2],
[3, 3, 4, 4]]
# tensor 't' is [[[1, 1, 1],
# [2, 2, 2]],
# [[3, 3, 3],
# [4, 4, 4]],
# [[5, 5, 5],
# [6, 6, 6]]]
# tensor 't' has shape [3, 2, 3]
# pass '[-1]' to flatten 't'
reshape(t, [-1]) ==> [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5, 6, 6, 6]
# -1 can also be used to infer the shape
# -1 is inferred to be 9:
reshape(t, [2, -1]) ==> [[1, 1, 1, 2, 2, 2, 3, 3, 3],
[4, 4, 4, 5, 5, 5, 6, 6, 6]]
# -1 is inferred to be 2:
reshape(t, [-1, 9]) ==> [[1, 1, 1, 2, 2, 2, 3, 3, 3],
[4, 4, 4, 5, 5, 5, 6, 6, 6]]
# -1 is inferred to be 3:
reshape(t, [ 2, -1, 3]) ==> [[[1, 1, 1],
[2, 2, 2],
[3, 3, 3]],
[[4, 4, 4],
[5, 5, 5],
[6, 6, 6]]]
# tensor 't' is [7]
# shape `[]` reshapes to a scalar
reshape(t, []) ==> 7
```
Args:
tensor: A `Tensor`.
shape: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Defines the shape of the output tensor.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Reshape", name,
tld.op_callbacks, tensor, shape)
return _result
except _core._FallbackException:
try:
return reshape_eager_fallback(
tensor, shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Reshape", tensor=tensor, shape=shape, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tshape",
_op._get_attr_type("Tshape"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Reshape", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Reshape = tf_export("raw_ops.Reshape")(_ops.to_raw_op(reshape))
def reshape_eager_fallback(tensor, shape, name, ctx):
_attr_T, (tensor,) = _execute.args_to_matching_eager([tensor], ctx)
_attr_Tshape, (shape,) = _execute.args_to_matching_eager([shape], ctx, _dtypes.int32)
_inputs_flat = [tensor, shape]
_attrs = ("T", _attr_T, "Tshape", _attr_Tshape)
_result = _execute.execute(b"Reshape", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Reshape", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def resource_strided_slice_assign(ref, begin, end, strides, value, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0, name=None):
r"""Assign `value` to the sliced l-value reference of `ref`.
The values of `value` are assigned to the positions in the variable
`ref` that are selected by the slice parameters. The slice parameters
`begin, `end`, `strides`, etc. work exactly as in `StridedSlice`.
NOTE this op currently does not support broadcasting and so `value`'s
shape must be exactly the shape produced by the slice of `ref`.
Args:
ref: A `Tensor` of type `resource`.
begin: A `Tensor`. Must be one of the following types: `int32`, `int64`.
end: A `Tensor`. Must have the same type as `begin`.
strides: A `Tensor`. Must have the same type as `begin`.
value: A `Tensor`.
begin_mask: An optional `int`. Defaults to `0`.
end_mask: An optional `int`. Defaults to `0`.
ellipsis_mask: An optional `int`. Defaults to `0`.
new_axis_mask: An optional `int`. Defaults to `0`.
shrink_axis_mask: An optional `int`. Defaults to `0`.
name: A name for the operation (optional).
Returns:
The created Operation.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ResourceStridedSliceAssign",
name, tld.op_callbacks, ref, begin, end, strides, value, "begin_mask",
begin_mask, "end_mask", end_mask, "ellipsis_mask", ellipsis_mask,
"new_axis_mask", new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
return _result
except _core._FallbackException:
try:
return resource_strided_slice_assign_eager_fallback(
ref, begin, end, strides, value, begin_mask=begin_mask,
end_mask=end_mask, ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask, shrink_axis_mask=shrink_axis_mask,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ResourceStridedSliceAssign", ref=ref, begin=begin, end=end,
strides=strides, value=value,
begin_mask=begin_mask,
end_mask=end_mask,
ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask,
shrink_axis_mask=shrink_axis_mask,
name=name)
return _op
ResourceStridedSliceAssign = tf_export("raw_ops.ResourceStridedSliceAssign")(_ops.to_raw_op(resource_strided_slice_assign))
def resource_strided_slice_assign_eager_fallback(ref, begin, end, strides, value, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask, name, ctx):
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_attr_T, (value,) = _execute.args_to_matching_eager([value], ctx)
_attr_Index, _inputs_Index = _execute.args_to_matching_eager([begin, end, strides], ctx)
(begin, end, strides) = _inputs_Index
ref = _ops.convert_to_tensor(ref, _dtypes.resource)
_inputs_flat = [ref, begin, end, strides, value]
_attrs = ("T", _attr_T, "Index", _attr_Index, "begin_mask", begin_mask,
"end_mask", end_mask, "ellipsis_mask", ellipsis_mask, "new_axis_mask",
new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
_result = _execute.execute(b"ResourceStridedSliceAssign", 0,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
_result = None
return _result
[文档]def reverse(tensor, dims, name=None):
r"""Reverses specific dimensions of a tensor.
Given a `tensor`, and a `bool` tensor `dims` representing the dimensions
of `tensor`, this operation reverses each dimension i of `tensor` where
`dims[i]` is `True`.
`tensor` can have up to 8 dimensions. The number of dimensions
of `tensor` must equal the number of elements in `dims`. In other words:
`rank(tensor) = size(dims)`
For example:
```
# tensor 't' is [[[[ 0, 1, 2, 3],
# [ 4, 5, 6, 7],
# [ 8, 9, 10, 11]],
# [[12, 13, 14, 15],
# [16, 17, 18, 19],
# [20, 21, 22, 23]]]]
# tensor 't' shape is [1, 2, 3, 4]
# 'dims' is [False, False, False, True]
reverse(t, dims) ==> [[[[ 3, 2, 1, 0],
[ 7, 6, 5, 4],
[ 11, 10, 9, 8]],
[[15, 14, 13, 12],
[19, 18, 17, 16],
[23, 22, 21, 20]]]]
# 'dims' is [False, True, False, False]
reverse(t, dims) ==> [[[[12, 13, 14, 15],
[16, 17, 18, 19],
[20, 21, 22, 23]
[[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11]]]]
# 'dims' is [False, False, True, False]
reverse(t, dims) ==> [[[[8, 9, 10, 11],
[4, 5, 6, 7],
[0, 1, 2, 3]]
[[20, 21, 22, 23],
[16, 17, 18, 19],
[12, 13, 14, 15]]]]
```
Args:
tensor: A `Tensor`. Must be one of the following types: `uint8`, `int8`, `uint16`, `int16`, `int32`, `int64`, `bool`, `half`, `float32`, `float64`, `complex64`, `complex128`, `string`.
Up to 8-D.
dims: A `Tensor` of type `bool`. 1-D. The dimensions to reverse.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Reverse", name,
tld.op_callbacks, tensor, dims)
return _result
except _core._FallbackException:
try:
return reverse_eager_fallback(
tensor, dims, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Reverse", tensor=tensor, dims=dims, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Reverse", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Reverse = tf_export("raw_ops.Reverse")(_ops.to_raw_op(reverse))
def reverse_eager_fallback(tensor, dims, name, ctx):
_attr_T, (tensor,) = _execute.args_to_matching_eager([tensor], ctx)
dims = _ops.convert_to_tensor(dims, _dtypes.bool)
_inputs_flat = [tensor, dims]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Reverse", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Reverse", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def reverse_sequence(input, seq_lengths, seq_dim, batch_dim=0, name=None):
r"""Reverses variable length slices.
This op first slices `input` along the dimension `batch_dim`, and for each
slice `i`, reverses the first `seq_lengths[i]` elements along
the dimension `seq_dim`.
The elements of `seq_lengths` must obey `seq_lengths[i] <= input.dims[seq_dim]`,
and `seq_lengths` must be a vector of length `input.dims[batch_dim]`.
The output slice `i` along dimension `batch_dim` is then given by input
slice `i`, with the first `seq_lengths[i]` slices along dimension
`seq_dim` reversed.
For example:
```
# Given this:
batch_dim = 0
seq_dim = 1
input.dims = (4, 8, ...)
seq_lengths = [7, 2, 3, 5]
# then slices of input are reversed on seq_dim, but only up to seq_lengths:
output[0, 0:7, :, ...] = input[0, 7:0:-1, :, ...]
output[1, 0:2, :, ...] = input[1, 2:0:-1, :, ...]
output[2, 0:3, :, ...] = input[2, 3:0:-1, :, ...]
output[3, 0:5, :, ...] = input[3, 5:0:-1, :, ...]
# while entries past seq_lens are copied through:
output[0, 7:, :, ...] = input[0, 7:, :, ...]
output[1, 2:, :, ...] = input[1, 2:, :, ...]
output[2, 3:, :, ...] = input[2, 3:, :, ...]
output[3, 2:, :, ...] = input[3, 2:, :, ...]
```
In contrast, if:
```
# Given this:
batch_dim = 2
seq_dim = 0
input.dims = (8, ?, 4, ...)
seq_lengths = [7, 2, 3, 5]
# then slices of input are reversed on seq_dim, but only up to seq_lengths:
output[0:7, :, 0, :, ...] = input[7:0:-1, :, 0, :, ...]
output[0:2, :, 1, :, ...] = input[2:0:-1, :, 1, :, ...]
output[0:3, :, 2, :, ...] = input[3:0:-1, :, 2, :, ...]
output[0:5, :, 3, :, ...] = input[5:0:-1, :, 3, :, ...]
# while entries past seq_lens are copied through:
output[7:, :, 0, :, ...] = input[7:, :, 0, :, ...]
output[2:, :, 1, :, ...] = input[2:, :, 1, :, ...]
output[3:, :, 2, :, ...] = input[3:, :, 2, :, ...]
output[2:, :, 3, :, ...] = input[2:, :, 3, :, ...]
```
Args:
input: A `Tensor`. The input to reverse.
seq_lengths: A `Tensor`. Must be one of the following types: `int32`, `int64`.
1-D with length `input.dims(batch_dim)` and
`max(seq_lengths) <= input.dims(seq_dim)`
seq_dim: An `int`. The dimension which is partially reversed.
batch_dim: An optional `int`. Defaults to `0`.
The dimension along which reversal is performed.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ReverseSequence", name,
tld.op_callbacks, input, seq_lengths, "seq_dim", seq_dim, "batch_dim",
batch_dim)
return _result
except _core._FallbackException:
try:
return reverse_sequence_eager_fallback(
input, seq_lengths, seq_dim=seq_dim, batch_dim=batch_dim,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
seq_dim = _execute.make_int(seq_dim, "seq_dim")
if batch_dim is None:
batch_dim = 0
batch_dim = _execute.make_int(batch_dim, "batch_dim")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ReverseSequence", input=input, seq_lengths=seq_lengths,
seq_dim=seq_dim, batch_dim=batch_dim, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("seq_dim", _op._get_attr_int("seq_dim"), "batch_dim",
_op._get_attr_int("batch_dim"), "T", _op._get_attr_type("T"),
"Tlen", _op._get_attr_type("Tlen"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ReverseSequence", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ReverseSequence = tf_export("raw_ops.ReverseSequence")(_ops.to_raw_op(reverse_sequence))
def reverse_sequence_eager_fallback(input, seq_lengths, seq_dim, batch_dim, name, ctx):
seq_dim = _execute.make_int(seq_dim, "seq_dim")
if batch_dim is None:
batch_dim = 0
batch_dim = _execute.make_int(batch_dim, "batch_dim")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tlen, (seq_lengths,) = _execute.args_to_matching_eager([seq_lengths], ctx, _dtypes.int64)
_inputs_flat = [input, seq_lengths]
_attrs = ("seq_dim", seq_dim, "batch_dim", batch_dim, "T", _attr_T, "Tlen",
_attr_Tlen)
_result = _execute.execute(b"ReverseSequence", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ReverseSequence", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('reverse', v1=['reverse', 'manip.reverse', 'reverse_v2'])
@deprecated_endpoints('manip.reverse', 'reverse_v2')
def reverse_v2(tensor, axis, name=None):
r"""Reverses specific dimensions of a tensor.
NOTE `tf.reverse` has now changed behavior in preparation for 1.0.
`tf.reverse_v2` is currently an alias that will be deprecated before TF 1.0.
Given a `tensor`, and a `int32` tensor `axis` representing the set of
dimensions of `tensor` to reverse. This operation reverses each dimension
`i` for which there exists `j` s.t. `axis[j] == i`.
`tensor` can have up to 8 dimensions. The number of dimensions specified
in `axis` may be 0 or more entries. If an index is specified more than
once, a InvalidArgument error is raised.
For example:
```
# tensor 't' is [[[[ 0, 1, 2, 3],
# [ 4, 5, 6, 7],
# [ 8, 9, 10, 11]],
# [[12, 13, 14, 15],
# [16, 17, 18, 19],
# [20, 21, 22, 23]]]]
# tensor 't' shape is [1, 2, 3, 4]
# 'dims' is [3] or 'dims' is [-1]
reverse(t, dims) ==> [[[[ 3, 2, 1, 0],
[ 7, 6, 5, 4],
[ 11, 10, 9, 8]],
[[15, 14, 13, 12],
[19, 18, 17, 16],
[23, 22, 21, 20]]]]
# 'dims' is '[1]' (or 'dims' is '[-3]')
reverse(t, dims) ==> [[[[12, 13, 14, 15],
[16, 17, 18, 19],
[20, 21, 22, 23]
[[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11]]]]
# 'dims' is '[2]' (or 'dims' is '[-2]')
reverse(t, dims) ==> [[[[8, 9, 10, 11],
[4, 5, 6, 7],
[0, 1, 2, 3]]
[[20, 21, 22, 23],
[16, 17, 18, 19],
[12, 13, 14, 15]]]]
```
Args:
tensor: A `Tensor`. Must be one of the following types: `uint8`, `int8`, `uint16`, `int16`, `int32`, `int64`, `bool`, `bfloat16`, `half`, `float32`, `float64`, `complex64`, `complex128`, `string`.
Up to 8-D.
axis: A `Tensor`. Must be one of the following types: `int32`, `int64`.
1-D. The indices of the dimensions to reverse. Must be in the range
`[-rank(tensor), rank(tensor))`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ReverseV2", name,
tld.op_callbacks, tensor, axis)
return _result
except _core._FallbackException:
try:
return reverse_v2_eager_fallback(
tensor, axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
reverse_v2, tensor=tensor, axis=axis, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ReverseV2", tensor=tensor, axis=axis, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
reverse_v2, tensor=tensor, axis=axis, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("Tidx", _op._get_attr_type("Tidx"), "T",
_op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ReverseV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ReverseV2 = tf_export("raw_ops.ReverseV2")(_ops.to_raw_op(reverse_v2))
def reverse_v2_eager_fallback(tensor, axis, name, ctx):
_attr_Tidx, (axis,) = _execute.args_to_matching_eager([axis], ctx, _dtypes.int32)
_attr_T, (tensor,) = _execute.args_to_matching_eager([tensor], ctx)
_inputs_flat = [tensor, axis]
_attrs = ("Tidx", _attr_Tidx, "T", _attr_T)
_result = _execute.execute(b"ReverseV2", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ReverseV2", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('scatter_nd', v1=['scatter_nd', 'manip.scatter_nd'])
@deprecated_endpoints('manip.scatter_nd')
def scatter_nd(indices, updates, shape, name=None):
r"""Scatter `updates` into a new tensor according to `indices`.
Creates a new tensor by applying sparse `updates` to individual values or
slices within a tensor (initially zero for numeric, empty for string) of
the given `shape` according to indices. This operator is the inverse of the
`tf.gather_nd` operator which extracts values or slices from a given tensor.
This operation is similar to tensor_scatter_add, except that the tensor is
zero-initialized. Calling `tf.scatter_nd(indices, values, shape)` is identical
to `tensor_scatter_add(tf.zeros(shape, values.dtype), indices, values)`
If `indices` contains duplicates, then their updates are accumulated (summed).
**WARNING**: The order in which updates are applied is nondeterministic, so the
output will be nondeterministic if `indices` contains duplicates -- because
of some numerical approximation issues, numbers summed in different order
may yield different results.
`indices` is an integer tensor containing indices into a new tensor of shape
`shape`. The last dimension of `indices` can be at most the rank of `shape`:
indices.shape[-1] <= shape.rank
The last dimension of `indices` corresponds to indices into elements
(if `indices.shape[-1] = shape.rank`) or slices
(if `indices.shape[-1] < shape.rank`) along dimension `indices.shape[-1]` of
`shape`. `updates` is a tensor with shape
indices.shape[:-1] + shape[indices.shape[-1]:]
The simplest form of scatter is to insert individual elements in a tensor by
index. For example, say we want to insert 4 scattered elements in a rank-1
tensor with 8 elements.
<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="https://www.tensorflow.org/images/ScatterNd1.png" alt>
</div>
In Python, this scatter operation would look like this:
```python
indices = tf.constant([[4], [3], [1], [7]])
updates = tf.constant([9, 10, 11, 12])
shape = tf.constant([8])
scatter = tf.scatter_nd(indices, updates, shape)
print(scatter)
```
The resulting tensor would look like this:
[0, 11, 0, 10, 9, 0, 0, 12]
We can also, insert entire slices of a higher rank tensor all at once. For
example, if we wanted to insert two slices in the first dimension of a
rank-3 tensor with two matrices of new values.
<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="https://www.tensorflow.org/images/ScatterNd2.png" alt>
</div>
In Python, this scatter operation would look like this:
```python
indices = tf.constant([[0], [2]])
updates = tf.constant([[[5, 5, 5, 5], [6, 6, 6, 6],
[7, 7, 7, 7], [8, 8, 8, 8]],
[[5, 5, 5, 5], [6, 6, 6, 6],
[7, 7, 7, 7], [8, 8, 8, 8]]])
shape = tf.constant([4, 4, 4])
scatter = tf.scatter_nd(indices, updates, shape)
print(scatter)
```
The resulting tensor would look like this:
[[[5, 5, 5, 5], [6, 6, 6, 6], [7, 7, 7, 7], [8, 8, 8, 8]],
[[0, 0, 0, 0], [0, 0, 0, 0], [0, 0, 0, 0], [0, 0, 0, 0]],
[[5, 5, 5, 5], [6, 6, 6, 6], [7, 7, 7, 7], [8, 8, 8, 8]],
[[0, 0, 0, 0], [0, 0, 0, 0], [0, 0, 0, 0], [0, 0, 0, 0]]]
Note that on CPU, if an out of bound index is found, an error is returned.
On GPU, if an out of bound index is found, the index is ignored.
Args:
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Index tensor.
updates: A `Tensor`. Updates to scatter into output.
shape: A `Tensor`. Must have the same type as `indices`.
1-D. The shape of the resulting tensor.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `updates`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ScatterNd", name,
tld.op_callbacks, indices, updates, shape)
return _result
except _core._FallbackException:
try:
return scatter_nd_eager_fallback(
indices, updates, shape, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
scatter_nd, indices=indices, updates=updates, shape=shape,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ScatterNd", indices=indices, updates=updates, shape=shape, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
scatter_nd, indices=indices, updates=updates, shape=shape,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ScatterNd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ScatterNd = tf_export("raw_ops.ScatterNd")(_ops.to_raw_op(scatter_nd))
def scatter_nd_eager_fallback(indices, updates, shape, name, ctx):
_attr_T, (updates,) = _execute.args_to_matching_eager([updates], ctx)
_attr_Tindices, _inputs_Tindices = _execute.args_to_matching_eager([indices, shape], ctx)
(indices, shape) = _inputs_Tindices
_inputs_flat = [indices, updates, shape]
_attrs = ("T", _attr_T, "Tindices", _attr_Tindices)
_result = _execute.execute(b"ScatterNd", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ScatterNd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def scatter_nd_non_aliasing_add(input, indices, updates, name=None):
r"""Applies sparse addition to `input` using individual values or slices
from `updates` according to indices `indices`. The updates are non-aliasing:
`input` is only modified in-place if no other operations will use it.
Otherwise, a copy of `input` is made. This operation has a gradient with
respect to both `input` and `updates`.
`input` is a `Tensor` with rank `P` and `indices` is a `Tensor` of rank `Q`.
`indices` must be integer tensor, containing indices into `input`.
It must be shape \\([d_0, ..., d_{Q-2}, K]\\) where `0 < K <= P`.
The innermost dimension of `indices` (with length `K`) corresponds to
indices into elements (if `K = P`) or `(P-K)`-dimensional slices
(if `K < P`) along the `K`th dimension of `input`.
`updates` is `Tensor` of rank `Q-1+P-K` with shape:
$$[d_0, ..., d_{Q-2}, input.shape[K], ..., input.shape[P-1]].$$
For example, say we want to add 4 scattered elements to a rank-1 tensor to 8
elements. In Python, that addition would look like this:
input = tf.constant([1, 2, 3, 4, 5, 6, 7, 8])
indices = tf.constant([[4], [3], [1], [7]])
updates = tf.constant([9, 10, 11, 12])
output = tf.scatter_nd_non_aliasing_add(input, indices, updates)
with tf.Session() as sess:
print(sess.run(output))
The resulting value `output` would look like this:
[1, 13, 3, 14, 14, 6, 7, 20]
See `tf.scatter_nd` for more details about how to make updates to slices.
Args:
input: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `uint8`, `int16`, `int8`, `complex64`, `int64`, `qint8`, `quint8`, `qint32`, `bfloat16`, `uint16`, `complex128`, `half`, `uint32`, `uint64`, `bool`.
A Tensor.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
A Tensor. Must be one of the following types: `int32`, `int64`.
A tensor of indices into `input`.
updates: A `Tensor`. Must have the same type as `input`.
A Tensor. Must have the same type as ref. A tensor of updated values
to add to `input`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ScatterNdNonAliasingAdd",
name, tld.op_callbacks, input, indices, updates)
return _result
except _core._FallbackException:
try:
return scatter_nd_non_aliasing_add_eager_fallback(
input, indices, updates, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ScatterNdNonAliasingAdd", input=input, indices=indices,
updates=updates, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ScatterNdNonAliasingAdd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ScatterNdNonAliasingAdd = tf_export("raw_ops.ScatterNdNonAliasingAdd")(_ops.to_raw_op(scatter_nd_non_aliasing_add))
def scatter_nd_non_aliasing_add_eager_fallback(input, indices, updates, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, updates], ctx)
(input, updates) = _inputs_T
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_inputs_flat = [input, indices, updates]
_attrs = ("T", _attr_T, "Tindices", _attr_Tindices)
_result = _execute.execute(b"ScatterNdNonAliasingAdd", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ScatterNdNonAliasingAdd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def shape(input, out_type=_dtypes.int32, name=None):
r"""Returns the shape of a tensor.
This operation returns a 1-D integer tensor representing the shape of `input`.
For example:
```
# 't' is [[[1, 1, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]]]
shape(t) ==> [2, 2, 3]
```
Args:
input: A `Tensor`.
out_type: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `out_type`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Shape", name,
tld.op_callbacks, input, "out_type", out_type)
return _result
except _core._FallbackException:
try:
return shape_eager_fallback(
input, out_type=out_type, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Shape", input=input, out_type=out_type, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_type",
_op._get_attr_type("out_type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Shape", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Shape = tf_export("raw_ops.Shape")(_ops.to_raw_op(shape))
def shape_eager_fallback(input, out_type, name, ctx):
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "out_type", out_type)
_result = _execute.execute(b"Shape", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Shape", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def shape_n(input, out_type=_dtypes.int32, name=None):
r"""Returns shape of tensors.
This operation returns N 1-D integer tensors representing shape of `input[i]s`.
Args:
input: A list of at least 1 `Tensor` objects with the same type.
out_type: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A list with the same length as `input` of `Tensor` objects with type `out_type`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ShapeN", name,
tld.op_callbacks, input, "out_type", out_type)
return _result
except _core._FallbackException:
try:
return shape_n_eager_fallback(
input, out_type=out_type, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if not isinstance(input, (list, tuple)):
raise TypeError(
"Expected list for 'input' argument to "
"'shape_n' Op, not %r." % input)
_attr_N = len(input)
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ShapeN", input=input, out_type=out_type, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("N", _op._get_attr_int("N"), "T", _op._get_attr_type("T"),
"out_type", _op._get_attr_type("out_type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ShapeN", _inputs_flat, _attrs, _result)
return _result
ShapeN = tf_export("raw_ops.ShapeN")(_ops.to_raw_op(shape_n))
def shape_n_eager_fallback(input, out_type, name, ctx):
if not isinstance(input, (list, tuple)):
raise TypeError(
"Expected list for 'input' argument to "
"'shape_n' Op, not %r." % input)
_attr_N = len(input)
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_attr_T, input = _execute.args_to_matching_eager(list(input), ctx)
_inputs_flat = list(input)
_attrs = ("N", _attr_N, "T", _attr_T, "out_type", out_type)
_result = _execute.execute(b"ShapeN", _attr_N, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ShapeN", _inputs_flat, _attrs, _result)
return _result
def size(input, out_type=_dtypes.int32, name=None):
r"""Returns the size of a tensor.
This operation returns an integer representing the number of elements in
`input`.
For example:
```
# 't' is [[[1, 1,, 1], [2, 2, 2]], [[3, 3, 3], [4, 4, 4]]]]
size(t) ==> 12
```
Args:
input: A `Tensor`.
out_type: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `out_type`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Size", name, tld.op_callbacks,
input, "out_type", out_type)
return _result
except _core._FallbackException:
try:
return size_eager_fallback(
input, out_type=out_type, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Size", input=input, out_type=out_type, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_type",
_op._get_attr_type("out_type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Size", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Size = tf_export("raw_ops.Size")(_ops.to_raw_op(size))
def size_eager_fallback(input, out_type, name, ctx):
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "out_type", out_type)
_result = _execute.execute(b"Size", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Size", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def _slice(input, begin, size, name=None):
r"""Return a slice from 'input'.
The output tensor is a tensor with dimensions described by 'size'
whose values are extracted from 'input' starting at the offsets in
'begin'.
*Requirements*:
0 <= begin[i] <= begin[i] + size[i] <= Di for i in [0, n)
Args:
input: A `Tensor`.
begin: A `Tensor`. Must be one of the following types: `int32`, `int64`.
begin[i] specifies the offset into the 'i'th dimension of
'input' to slice from.
size: A `Tensor`. Must have the same type as `begin`.
size[i] specifies the number of elements of the 'i'th dimension
of 'input' to slice. If size[i] is -1, all remaining elements in dimension
i are included in the slice (i.e. this is equivalent to setting
size[i] = input.dim_size(i) - begin[i]).
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Slice", name,
tld.op_callbacks, input, begin, size)
return _result
except _core._FallbackException:
try:
return _slice_eager_fallback(
input, begin, size, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Slice", input=input, begin=begin, size=size, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Index",
_op._get_attr_type("Index"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Slice", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Slice = tf_export("raw_ops.Slice")(_ops.to_raw_op(_slice))
def _slice_eager_fallback(input, begin, size, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Index, _inputs_Index = _execute.args_to_matching_eager([begin, size], ctx)
(begin, size) = _inputs_Index
_inputs_flat = [input, begin, size]
_attrs = ("T", _attr_T, "Index", _attr_Index)
_result = _execute.execute(b"Slice", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Slice", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def snapshot(input, name=None):
r"""Returns a copy of the input tensor.
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Snapshot", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return snapshot_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Snapshot", input=input, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Snapshot", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Snapshot = tf_export("raw_ops.Snapshot")(_ops.to_raw_op(snapshot))
def snapshot_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Snapshot", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Snapshot", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def space_to_batch(input, paddings, block_size, name=None):
r"""SpaceToBatch for 4-D tensors of type T.
This is a legacy version of the more general SpaceToBatchND.
Zero-pads and then rearranges (permutes) blocks of spatial data into batch.
More specifically, this op outputs a copy of the input tensor where values from
the `height` and `width` dimensions are moved to the `batch` dimension. After
the zero-padding, both `height` and `width` of the input must be divisible by the
block size.
Args:
input: A `Tensor`. 4-D with shape `[batch, height, width, depth]`.
paddings: A `Tensor`. Must be one of the following types: `int32`, `int64`.
2-D tensor of non-negative integers with shape `[2, 2]`. It specifies
the padding of the input with zeros across the spatial dimensions as follows:
paddings = [[pad_top, pad_bottom], [pad_left, pad_right]]
The effective spatial dimensions of the zero-padded input tensor will be:
height_pad = pad_top + height + pad_bottom
width_pad = pad_left + width + pad_right
The attr `block_size` must be greater than one. It indicates the block size.
* Non-overlapping blocks of size `block_size x block size` in the height and
width dimensions are rearranged into the batch dimension at each location.
* The batch of the output tensor is `batch * block_size * block_size`.
* Both height_pad and width_pad must be divisible by block_size.
The shape of the output will be:
[batch*block_size*block_size, height_pad/block_size, width_pad/block_size,
depth]
Some examples:
(1) For the following input of shape `[1, 2, 2, 1]` and block_size of 2:
```
x = [[[[1], [2]], [[3], [4]]]]
```
The output tensor has shape `[4, 1, 1, 1]` and value:
```
[[[[1]]], [[[2]]], [[[3]]], [[[4]]]]
```
(2) For the following input of shape `[1, 2, 2, 3]` and block_size of 2:
```
x = [[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
```
The output tensor has shape `[4, 1, 1, 3]` and value:
```
[[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]]
```
(3) For the following input of shape `[1, 4, 4, 1]` and block_size of 2:
```
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]],
[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
```
The output tensor has shape `[4, 2, 2, 1]` and value:
```
x = [[[[1], [3]], [[9], [11]]],
[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]
```
(4) For the following input of shape `[2, 2, 4, 1]` and block_size of 2:
```
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]]],
[[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
```
The output tensor has shape `[8, 1, 2, 1]` and value:
```
x = [[[[1], [3]]], [[[9], [11]]], [[[2], [4]]], [[[10], [12]]],
[[[5], [7]]], [[[13], [15]]], [[[6], [8]]], [[[14], [16]]]]
```
Among others, this operation is useful for reducing atrous convolution into
regular convolution.
block_size: An `int` that is `>= 2`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "SpaceToBatch", name,
tld.op_callbacks, input, paddings, "block_size", block_size)
return _result
except _core._FallbackException:
try:
return space_to_batch_eager_fallback(
input, paddings, block_size=block_size, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
block_size = _execute.make_int(block_size, "block_size")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"SpaceToBatch", input=input, paddings=paddings, block_size=block_size,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tpaddings",
_op._get_attr_type("Tpaddings"), "block_size",
_op._get_attr_int("block_size"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"SpaceToBatch", _inputs_flat, _attrs, _result)
_result, = _result
return _result
SpaceToBatch = tf_export("raw_ops.SpaceToBatch")(_ops.to_raw_op(space_to_batch))
def space_to_batch_eager_fallback(input, paddings, block_size, name, ctx):
block_size = _execute.make_int(block_size, "block_size")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tpaddings, (paddings,) = _execute.args_to_matching_eager([paddings], ctx, _dtypes.int32)
_inputs_flat = [input, paddings]
_attrs = ("T", _attr_T, "Tpaddings", _attr_Tpaddings, "block_size",
block_size)
_result = _execute.execute(b"SpaceToBatch", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"SpaceToBatch", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('space_to_batch_nd', v1=['space_to_batch_nd', 'manip.space_to_batch_nd'])
@deprecated_endpoints('manip.space_to_batch_nd')
def space_to_batch_nd(input, block_shape, paddings, name=None):
r"""SpaceToBatch for N-D tensors of type T.
This operation divides "spatial" dimensions `[1, ..., M]` of the input into a
grid of blocks of shape `block_shape`, and interleaves these blocks with the
"batch" dimension (0) such that in the output, the spatial dimensions
`[1, ..., M]` correspond to the position within the grid, and the batch
dimension combines both the position within a spatial block and the original
batch position. Prior to division into blocks, the spatial dimensions of the
input are optionally zero padded according to `paddings`. See below for a
precise description.
Args:
input: A `Tensor`.
N-D with shape `input_shape = [batch] + spatial_shape + remaining_shape`,
where spatial_shape has `M` dimensions.
block_shape: A `Tensor`. Must be one of the following types: `int32`, `int64`.
1-D with shape `[M]`, all values must be >= 1.
paddings: A `Tensor`. Must be one of the following types: `int32`, `int64`.
2-D with shape `[M, 2]`, all values must be >= 0.
`paddings[i] = [pad_start, pad_end]` specifies the padding for input dimension
`i + 1`, which corresponds to spatial dimension `i`. It is required that
`block_shape[i]` divides `input_shape[i + 1] + pad_start + pad_end`.
This operation is equivalent to the following steps:
1. Zero-pad the start and end of dimensions `[1, ..., M]` of the
input according to `paddings` to produce `padded` of shape `padded_shape`.
2. Reshape `padded` to `reshaped_padded` of shape:
[batch] +
[padded_shape[1] / block_shape[0],
block_shape[0],
...,
padded_shape[M] / block_shape[M-1],
block_shape[M-1]] +
remaining_shape
3. Permute dimensions of `reshaped_padded` to produce
`permuted_reshaped_padded` of shape:
block_shape +
[batch] +
[padded_shape[1] / block_shape[0],
...,
padded_shape[M] / block_shape[M-1]] +
remaining_shape
4. Reshape `permuted_reshaped_padded` to flatten `block_shape` into the batch
dimension, producing an output tensor of shape:
[batch * prod(block_shape)] +
[padded_shape[1] / block_shape[0],
...,
padded_shape[M] / block_shape[M-1]] +
remaining_shape
Some examples:
(1) For the following input of shape `[1, 2, 2, 1]`, `block_shape = [2, 2]`, and
`paddings = [[0, 0], [0, 0]]`:
```
x = [[[[1], [2]], [[3], [4]]]]
```
The output tensor has shape `[4, 1, 1, 1]` and value:
```
[[[[1]]], [[[2]]], [[[3]]], [[[4]]]]
```
(2) For the following input of shape `[1, 2, 2, 3]`, `block_shape = [2, 2]`, and
`paddings = [[0, 0], [0, 0]]`:
```
x = [[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
```
The output tensor has shape `[4, 1, 1, 3]` and value:
```
[[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]]
```
(3) For the following input of shape `[1, 4, 4, 1]`, `block_shape = [2, 2]`, and
`paddings = [[0, 0], [0, 0]]`:
```
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]],
[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
```
The output tensor has shape `[4, 2, 2, 1]` and value:
```
x = [[[[1], [3]], [[9], [11]]],
[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]
```
(4) For the following input of shape `[2, 2, 4, 1]`, block_shape = `[2, 2]`, and
paddings = `[[0, 0], [2, 0]]`:
```
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]]],
[[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
```
The output tensor has shape `[8, 1, 3, 1]` and value:
```
x = [[[[0], [1], [3]]], [[[0], [9], [11]]],
[[[0], [2], [4]]], [[[0], [10], [12]]],
[[[0], [5], [7]]], [[[0], [13], [15]]],
[[[0], [6], [8]]], [[[0], [14], [16]]]]
```
Among others, this operation is useful for reducing atrous convolution into
regular convolution.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "SpaceToBatchND", name,
tld.op_callbacks, input, block_shape, paddings)
return _result
except _core._FallbackException:
try:
return space_to_batch_nd_eager_fallback(
input, block_shape, paddings, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
space_to_batch_nd, input=input, block_shape=block_shape,
paddings=paddings, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"SpaceToBatchND", input=input, block_shape=block_shape,
paddings=paddings, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
space_to_batch_nd, input=input, block_shape=block_shape,
paddings=paddings, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tblock_shape",
_op._get_attr_type("Tblock_shape"), "Tpaddings",
_op._get_attr_type("Tpaddings"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"SpaceToBatchND", _inputs_flat, _attrs, _result)
_result, = _result
return _result
SpaceToBatchND = tf_export("raw_ops.SpaceToBatchND")(_ops.to_raw_op(space_to_batch_nd))
def space_to_batch_nd_eager_fallback(input, block_shape, paddings, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tblock_shape, (block_shape,) = _execute.args_to_matching_eager([block_shape], ctx, _dtypes.int32)
_attr_Tpaddings, (paddings,) = _execute.args_to_matching_eager([paddings], ctx, _dtypes.int32)
_inputs_flat = [input, block_shape, paddings]
_attrs = ("T", _attr_T, "Tblock_shape", _attr_Tblock_shape, "Tpaddings",
_attr_Tpaddings)
_result = _execute.execute(b"SpaceToBatchND", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"SpaceToBatchND", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def space_to_depth(input, block_size, data_format="NHWC", name=None):
r"""SpaceToDepth for tensors of type T.
Rearranges blocks of spatial data, into depth. More specifically,
this op outputs a copy of the input tensor where values from the `height`
and `width` dimensions are moved to the `depth` dimension.
The attr `block_size` indicates the input block size.
* Non-overlapping blocks of size `block_size x block size` are rearranged
into depth at each location.
* The depth of the output tensor is `block_size * block_size * input_depth`.
* The Y, X coordinates within each block of the input become the high order
component of the output channel index.
* The input tensor's height and width must be divisible by block_size.
The `data_format` attr specifies the layout of the input and output tensors
with the following options:
"NHWC": `[ batch, height, width, channels ]`
"NCHW": `[ batch, channels, height, width ]`
"NCHW_VECT_C":
`qint8 [ batch, channels / 4, height, width, 4 ]`
It is useful to consider the operation as transforming a 6-D Tensor.
e.g. for data_format = NHWC,
Each element in the input tensor can be specified via 6 coordinates,
ordered by decreasing memory layout significance as:
n,oY,bY,oX,bX,iC (where n=batch index, oX, oY means X or Y coordinates
within the output image, bX, bY means coordinates
within the input block, iC means input channels).
The output would be a transpose to the following layout:
n,oY,oX,bY,bX,iC
This operation is useful for resizing the activations between convolutions
(but keeping all data), e.g. instead of pooling. It is also useful for training
purely convolutional models.
For example, given an input of shape `[1, 2, 2, 1]`, data_format = "NHWC" and
block_size = 2:
```
x = [[[[1], [2]],
[[3], [4]]]]
```
This operation will output a tensor of shape `[1, 1, 1, 4]`:
```
[[[[1, 2, 3, 4]]]]
```
Here, the input has a batch of 1 and each batch element has shape `[2, 2, 1]`,
the corresponding output will have a single element (i.e. width and height are
both 1) and will have a depth of 4 channels (1 * block_size * block_size).
The output element shape is `[1, 1, 4]`.
For an input tensor with larger depth, here of shape `[1, 2, 2, 3]`, e.g.
```
x = [[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
```
This operation, for block_size of 2, will return the following tensor of shape
`[1, 1, 1, 12]`
```
[[[[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]]]]
```
Similarly, for the following input of shape `[1 4 4 1]`, and a block size of 2:
```
x = [[[[1], [2], [5], [6]],
[[3], [4], [7], [8]],
[[9], [10], [13], [14]],
[[11], [12], [15], [16]]]]
```
the operator will return the following tensor of shape `[1 2 2 4]`:
```
x = [[[[1, 2, 3, 4],
[5, 6, 7, 8]],
[[9, 10, 11, 12],
[13, 14, 15, 16]]]]
```
Args:
input: A `Tensor`.
block_size: An `int` that is `>= 2`. The size of the spatial block.
data_format: An optional `string` from: `"NHWC", "NCHW", "NCHW_VECT_C"`. Defaults to `"NHWC"`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "SpaceToDepth", name,
tld.op_callbacks, input, "block_size", block_size, "data_format",
data_format)
return _result
except _core._FallbackException:
try:
return space_to_depth_eager_fallback(
input, block_size=block_size, data_format=data_format, name=name,
ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
block_size = _execute.make_int(block_size, "block_size")
if data_format is None:
data_format = "NHWC"
data_format = _execute.make_str(data_format, "data_format")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"SpaceToDepth", input=input, block_size=block_size,
data_format=data_format, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "block_size",
_op._get_attr_int("block_size"), "data_format",
_op.get_attr("data_format"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"SpaceToDepth", _inputs_flat, _attrs, _result)
_result, = _result
return _result
SpaceToDepth = tf_export("raw_ops.SpaceToDepth")(_ops.to_raw_op(space_to_depth))
def space_to_depth_eager_fallback(input, block_size, data_format, name, ctx):
block_size = _execute.make_int(block_size, "block_size")
if data_format is None:
data_format = "NHWC"
data_format = _execute.make_str(data_format, "data_format")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "block_size", block_size, "data_format",
data_format)
_result = _execute.execute(b"SpaceToDepth", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"SpaceToDepth", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def split(axis, value, num_split, name=None):
r"""Splits a tensor into `num_split` tensors along one dimension.
Args:
axis: A `Tensor` of type `int32`.
0-D. The dimension along which to split. Must be in the range
`[-rank(value), rank(value))`.
value: A `Tensor`. The tensor to split.
num_split: An `int` that is `>= 1`.
The number of ways to split. Must evenly divide
`value.shape[split_dim]`.
name: A name for the operation (optional).
Returns:
A list of `num_split` `Tensor` objects with the same type as `value`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Split", name,
tld.op_callbacks, axis, value, "num_split", num_split)
return _result
except _core._FallbackException:
try:
return split_eager_fallback(
axis, value, num_split=num_split, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
num_split = _execute.make_int(num_split, "num_split")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Split", split_dim=axis, value=value, num_split=num_split, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num_split", _op._get_attr_int("num_split"), "T",
_op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Split", _inputs_flat, _attrs, _result)
return _result
Split = tf_export("raw_ops.Split")(_ops.to_raw_op(split))
def split_eager_fallback(axis, value, num_split, name, ctx):
num_split = _execute.make_int(num_split, "num_split")
_attr_T, (value,) = _execute.args_to_matching_eager([value], ctx)
axis = _ops.convert_to_tensor(axis, _dtypes.int32)
_inputs_flat = [axis, value]
_attrs = ("num_split", num_split, "T", _attr_T)
_result = _execute.execute(b"Split", num_split, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Split", _inputs_flat, _attrs, _result)
return _result
def split_v(value, size_splits, axis, num_split, name=None):
r"""Splits a tensor into `num_split` tensors along one dimension.
Args:
value: A `Tensor`. The tensor to split.
size_splits: A `Tensor`. Must be one of the following types: `int32`, `int64`.
list containing the sizes of each output tensor along the split
dimension. Must sum to the dimension of value along split_dim.
Can contain one -1 indicating that dimension is to be inferred.
axis: A `Tensor` of type `int32`.
0-D. The dimension along which to split. Must be in the range
`[-rank(value), rank(value))`.
num_split: An `int` that is `>= 1`.
name: A name for the operation (optional).
Returns:
A list of `num_split` `Tensor` objects with the same type as `value`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "SplitV", name,
tld.op_callbacks, value, size_splits, axis, "num_split", num_split)
return _result
except _core._FallbackException:
try:
return split_v_eager_fallback(
value, size_splits, axis, num_split=num_split, name=name,
ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
num_split = _execute.make_int(num_split, "num_split")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"SplitV", value=value, size_splits=size_splits, split_dim=axis,
num_split=num_split, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num_split", _op._get_attr_int("num_split"), "T",
_op._get_attr_type("T"), "Tlen", _op._get_attr_type("Tlen"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"SplitV", _inputs_flat, _attrs, _result)
return _result
SplitV = tf_export("raw_ops.SplitV")(_ops.to_raw_op(split_v))
def split_v_eager_fallback(value, size_splits, axis, num_split, name, ctx):
num_split = _execute.make_int(num_split, "num_split")
_attr_T, (value,) = _execute.args_to_matching_eager([value], ctx)
_attr_Tlen, (size_splits,) = _execute.args_to_matching_eager([size_splits], ctx, _dtypes.int64)
axis = _ops.convert_to_tensor(axis, _dtypes.int32)
_inputs_flat = [value, size_splits, axis]
_attrs = ("num_split", num_split, "T", _attr_T, "Tlen", _attr_Tlen)
_result = _execute.execute(b"SplitV", num_split, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"SplitV", _inputs_flat, _attrs, _result)
return _result
def squeeze(input, axis=[], name=None):
r"""Removes dimensions of size 1 from the shape of a tensor.
Given a tensor `input`, this operation returns a tensor of the same type with
all dimensions of size 1 removed. If you don't want to remove all size 1
dimensions, you can remove specific size 1 dimensions by specifying
`axis`.
For example:
```
# 't' is a tensor of shape [1, 2, 1, 3, 1, 1]
shape(squeeze(t)) ==> [2, 3]
```
Or, to remove specific size 1 dimensions:
```
# 't' is a tensor of shape [1, 2, 1, 3, 1, 1]
shape(squeeze(t, [2, 4])) ==> [1, 2, 3, 1]
```
Args:
input: A `Tensor`. The `input` to squeeze.
axis: An optional list of `ints`. Defaults to `[]`.
If specified, only squeezes the dimensions listed. The dimension
index starts at 0. It is an error to squeeze a dimension that is not 1. Must
be in the range `[-rank(input), rank(input))`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Squeeze", name,
tld.op_callbacks, input, "squeeze_dims", axis)
return _result
except _core._FallbackException:
try:
return squeeze_eager_fallback(
input, axis=axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if axis is None:
axis = []
if not isinstance(axis, (list, tuple)):
raise TypeError(
"Expected list for 'axis' argument to "
"'squeeze' Op, not %r." % axis)
axis = [_execute.make_int(_i, "axis") for _i in axis]
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Squeeze", input=input, squeeze_dims=axis, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "squeeze_dims",
_op.get_attr("squeeze_dims"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Squeeze", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Squeeze = tf_export("raw_ops.Squeeze")(_ops.to_raw_op(squeeze))
def squeeze_eager_fallback(input, axis, name, ctx):
if axis is None:
axis = []
if not isinstance(axis, (list, tuple)):
raise TypeError(
"Expected list for 'axis' argument to "
"'squeeze' Op, not %r." % axis)
axis = [_execute.make_int(_i, "axis") for _i in axis]
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T, "squeeze_dims", axis)
_result = _execute.execute(b"Squeeze", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Squeeze", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('stop_gradient')
def stop_gradient(input, name=None):
r"""Stops gradient computation.
When executed in a graph, this op outputs its input tensor as-is.
When building ops to compute gradients, this op prevents the contribution of
its inputs to be taken into account. Normally, the gradient generator adds ops
to a graph to compute the derivatives of a specified 'loss' by recursively
finding out inputs that contributed to its computation. If you insert this op
in the graph it inputs are masked from the gradient generator. They are not
taken into account for computing gradients.
This is useful any time you want to compute a value with TensorFlow but need
to pretend that the value was a constant. Some examples include:
* The *EM* algorithm where the *M-step* should not involve backpropagation
through the output of the *E-step*.
* Contrastive divergence training of Boltzmann machines where, when
differentiating the energy function, the training must not backpropagate
through the graph that generated the samples from the model.
* Adversarial training, where no backprop should happen through the adversarial
example generation process.
Args:
input: A `Tensor`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "StopGradient", name,
tld.op_callbacks, input)
return _result
except _core._FallbackException:
try:
return stop_gradient_eager_fallback(
input, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
stop_gradient, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"StopGradient", input=input, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
stop_gradient, input=input, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"StopGradient", _inputs_flat, _attrs, _result)
_result, = _result
return _result
StopGradient = tf_export("raw_ops.StopGradient")(_ops.to_raw_op(stop_gradient))
def stop_gradient_eager_fallback(input, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_inputs_flat = [input]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"StopGradient", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"StopGradient", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def strided_slice(input, begin, end, strides, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0, name=None):
r"""Return a strided slice from `input`.
Note, most python users will want to use the Python `Tensor.__getitem__`
or `Variable.__getitem__` rather than this op directly.
The goal of this op is to produce a new tensor with a subset of
the elements from the `n` dimensional `input` tensor. The subset is chosen using
a sequence of `m` sparse range specifications encoded into the arguments
of this function. Note, in some cases
`m` could be equal to `n`, but this need not be the case. Each
range specification entry can be one of the following:
- An ellipsis (...). Ellipses are used to imply zero or more
dimensions of full-dimension selection and are produced using
`ellipsis_mask`. For example, `foo[...]` is the identity slice.
- A new axis. This is used to insert a new shape=1 dimension and is
produced using `new_axis_mask`. For example, `foo[:, ...]` where
`foo` is shape `(3, 4)` produces a `(1, 3, 4)` tensor.
- A range `begin:end:stride`. This is used to specify how much to choose from
a given dimension. `stride` can be any integer but 0. `begin` is an integer
which represents the index of the first value to select while `end` represents
the index of the last value to select. The number of values selected in each
dimension is `end - begin` if `stride > 0` and `begin - end` if `stride < 0`.
`begin` and `end` can be negative where `-1` is the last element, `-2` is
the second to last. `begin_mask` controls whether to replace the explicitly
given `begin` with an implicit effective value of `0` if `stride > 0` and
`-1` if `stride < 0`. `end_mask` is analogous but produces the number
required to create the largest open interval. For example, given a shape
`(3,)` tensor `foo[:]`, the effective `begin` and `end` are `0` and `3`. Do
not assume this is equivalent to `foo[0:-1]` which has an effective `begin`
and `end` of `0` and `2`. Another example is `foo[-2::-1]` which reverses the
first dimension of a tensor while dropping the last two (in the original
order elements). For example `foo = [1,2,3,4]; foo[-2::-1]` is `[4,3]`.
- A single index. This is used to keep only elements that have a given
index. For example (`foo[2, :]` on a shape `(5,6)` tensor produces a
shape `(6,)` tensor. This is encoded in `begin` and `end` and
`shrink_axis_mask`.
Each conceptual range specification is encoded in the op's argument. This
encoding is best understand by considering a non-trivial example. In
particular,
`foo[1, 2:4, None, ..., :-3:-1, :]` will be encoded as
```
begin = [1, 2, x, x, 0, x] # x denotes don't care (usually 0)
end = [2, 4, x, x, -3, x]
strides = [1, 1, x, x, -1, 1]
begin_mask = 1<<4 | 1 << 5 = 48
end_mask = 1<<5 = 32
ellipsis_mask = 1<<3 = 8
new_axis_mask = 1<<2 4
shrink_axis_mask = 1<<0
```
In this case if `foo.shape` is (5, 5, 5, 5, 5, 5) the final shape of
the slice becomes (2, 1, 5, 5, 2, 5).
Let us walk step by step through each argument specification.
1. The first argument in the example slice is turned into `begin = 1` and
`end = begin + 1 = 2`. To disambiguate from the original spec `2:4` we
also set the appropriate bit in `shrink_axis_mask`.
2. `2:4` is contributes 2, 4, 1 to begin, end, and stride. All masks have
zero bits contributed.
3. None is a synonym for `tf.newaxis`. This means insert a dimension of size 1
dimension in the final shape. Dummy values are contributed to begin,
end and stride, while the new_axis_mask bit is set.
4. `...` grab the full ranges from as many dimensions as needed to
fully specify a slice for every dimension of the input shape.
5. `:-3:-1` shows the use of negative indices. A negative index `i` associated
with a dimension that has shape `s` is converted to a positive index
`s + i`. So `-1` becomes `s-1` (i.e. the last element). This conversion
is done internally so begin, end and strides receive x, -3, and -1.
The appropriate begin_mask bit is set to indicate the start range is the
full range (ignoring the x).
6. `:` indicates that the entire contents of the corresponding dimension
is selected. This is equivalent to `::` or `0::1`. begin, end, and strides
receive 0, 0, and 1, respectively. The appropriate bits in `begin_mask` and
`end_mask` are also set.
*Requirements*:
`0 != strides[i] for i in [0, m)`
`ellipsis_mask must be a power of two (only one ellipsis)`
Args:
input: A `Tensor`.
begin: A `Tensor`. Must be one of the following types: `int32`, `int64`.
`begin[k]` specifies the offset into the `k`th range specification.
The exact dimension this corresponds to will be determined by context.
Out-of-bounds values will be silently clamped. If the `k`th bit of
`begin_mask` then `begin[k]` is ignored and the full range of the
appropriate dimension is used instead. Negative values causes indexing
to start from the highest element e.g. If `foo==[1,2,3]` then `foo[-1]==3`.
end: A `Tensor`. Must have the same type as `begin`.
`end[i]` is like `begin` with the exception that `end_mask` is
used to determine full ranges.
strides: A `Tensor`. Must have the same type as `begin`.
`strides[i]` specifies the increment in the `i`th specification
after extracting a given element. Negative indices will reverse
the original order. Out or range values are
clamped to `[0,dim[i]) if slice[i]>0` or `[-1,dim[i]-1] if slice[i] < 0`
begin_mask: An optional `int`. Defaults to `0`.
a bitmask where a bit i being 1 means to ignore the begin
value and instead use the largest interval possible. At runtime
begin[i] will be replaced with `[0, n-1)` if `stride[i] > 0` or
`[-1, n-1]` if `stride[i] < 0`
end_mask: An optional `int`. Defaults to `0`. analogous to `begin_mask`
ellipsis_mask: An optional `int`. Defaults to `0`.
a bitmask where bit `i` being 1 means the `i`th
position is actually an ellipsis. One bit at most can be 1.
If `ellipsis_mask == 0`, then an implicit ellipsis mask of `1 << (m+1)`
is provided. This means that `foo[3:5] == foo[3:5, ...]`. An ellipsis
implicitly creates as many range specifications as necessary to fully
specify the sliced range for every dimension. For example for a 4-dimensional
tensor `foo` the slice `foo[2, ..., 5:8]` implies `foo[2, :, :, 5:8]`.
new_axis_mask: An optional `int`. Defaults to `0`.
a bitmask where bit `i` being 1 means the `i`th
specification creates a new shape 1 dimension. For example
`foo[:4, tf.newaxis, :2]` would produce a shape `(4, 1, 2)` tensor.
shrink_axis_mask: An optional `int`. Defaults to `0`.
a bitmask where bit `i` implies that the `i`th
specification should shrink the dimensionality. begin and end
must imply a slice of size 1 in the dimension. For example in
python one might do `foo[:, 3, :]` which would result in
`shrink_axis_mask` being 2.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "StridedSlice", name,
tld.op_callbacks, input, begin, end, strides, "begin_mask",
begin_mask, "end_mask", end_mask, "ellipsis_mask", ellipsis_mask,
"new_axis_mask", new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
return _result
except _core._FallbackException:
try:
return strided_slice_eager_fallback(
input, begin, end, strides, begin_mask=begin_mask,
end_mask=end_mask, ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask, shrink_axis_mask=shrink_axis_mask,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"StridedSlice", input=input, begin=begin, end=end, strides=strides,
begin_mask=begin_mask, end_mask=end_mask,
ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask,
shrink_axis_mask=shrink_axis_mask, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Index",
_op._get_attr_type("Index"), "begin_mask",
_op._get_attr_int("begin_mask"), "end_mask",
_op._get_attr_int("end_mask"), "ellipsis_mask",
_op._get_attr_int("ellipsis_mask"), "new_axis_mask",
_op._get_attr_int("new_axis_mask"), "shrink_axis_mask",
_op._get_attr_int("shrink_axis_mask"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"StridedSlice", _inputs_flat, _attrs, _result)
_result, = _result
return _result
StridedSlice = tf_export("raw_ops.StridedSlice")(_ops.to_raw_op(strided_slice))
def strided_slice_eager_fallback(input, begin, end, strides, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask, name, ctx):
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Index, _inputs_Index = _execute.args_to_matching_eager([begin, end, strides], ctx)
(begin, end, strides) = _inputs_Index
_inputs_flat = [input, begin, end, strides]
_attrs = ("T", _attr_T, "Index", _attr_Index, "begin_mask", begin_mask,
"end_mask", end_mask, "ellipsis_mask", ellipsis_mask, "new_axis_mask",
new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
_result = _execute.execute(b"StridedSlice", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"StridedSlice", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def strided_slice_assign(ref, begin, end, strides, value, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0, name=None):
r"""Assign `value` to the sliced l-value reference of `ref`.
The values of `value` are assigned to the positions in the variable
`ref` that are selected by the slice parameters. The slice parameters
`begin`, `end`, `strides`, etc. work exactly as in `StridedSlice`.
NOTE this op currently does not support broadcasting and so `value`'s
shape must be exactly the shape produced by the slice of `ref`.
Args:
ref: A mutable `Tensor`.
begin: A `Tensor`. Must be one of the following types: `int32`, `int64`.
end: A `Tensor`. Must have the same type as `begin`.
strides: A `Tensor`. Must have the same type as `begin`.
value: A `Tensor`. Must have the same type as `ref`.
begin_mask: An optional `int`. Defaults to `0`.
end_mask: An optional `int`. Defaults to `0`.
ellipsis_mask: An optional `int`. Defaults to `0`.
new_axis_mask: An optional `int`. Defaults to `0`.
shrink_axis_mask: An optional `int`. Defaults to `0`.
name: A name for the operation (optional).
Returns:
A mutable `Tensor`. Has the same type as `ref`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
raise RuntimeError("strided_slice_assign op does not support eager execution. Arg 'output_ref' is a ref.")
# Add nodes to the TensorFlow graph.
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"StridedSliceAssign", ref=ref, begin=begin, end=end, strides=strides,
value=value, begin_mask=begin_mask,
end_mask=end_mask, ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask,
shrink_axis_mask=shrink_axis_mask, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Index",
_op._get_attr_type("Index"), "begin_mask",
_op._get_attr_int("begin_mask"), "end_mask",
_op._get_attr_int("end_mask"), "ellipsis_mask",
_op._get_attr_int("ellipsis_mask"), "new_axis_mask",
_op._get_attr_int("new_axis_mask"), "shrink_axis_mask",
_op._get_attr_int("shrink_axis_mask"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"StridedSliceAssign", _inputs_flat, _attrs, _result)
_result, = _result
return _result
StridedSliceAssign = tf_export("raw_ops.StridedSliceAssign")(_ops.to_raw_op(strided_slice_assign))
def strided_slice_assign_eager_fallback(ref, begin, end, strides, value, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask, name, ctx):
raise RuntimeError("strided_slice_assign op does not support eager execution. Arg 'output_ref' is a ref.")
def strided_slice_grad(shape, begin, end, strides, dy, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0, name=None):
r"""Returns the gradient of `StridedSlice`.
Since `StridedSlice` cuts out pieces of its `input` which is size
`shape`, its gradient will have the same shape (which is passed here
as `shape`). The gradient will be zero in any element that the slice
does not select.
Arguments are the same as StridedSliceGrad with the exception that
`dy` is the input gradient to be propagated and `shape` is the
shape of `StridedSlice`'s `input`.
Args:
shape: A `Tensor`. Must be one of the following types: `int32`, `int64`.
begin: A `Tensor`. Must have the same type as `shape`.
end: A `Tensor`. Must have the same type as `shape`.
strides: A `Tensor`. Must have the same type as `shape`.
dy: A `Tensor`.
begin_mask: An optional `int`. Defaults to `0`.
end_mask: An optional `int`. Defaults to `0`.
ellipsis_mask: An optional `int`. Defaults to `0`.
new_axis_mask: An optional `int`. Defaults to `0`.
shrink_axis_mask: An optional `int`. Defaults to `0`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `dy`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "StridedSliceGrad", name,
tld.op_callbacks, shape, begin, end, strides, dy, "begin_mask",
begin_mask, "end_mask", end_mask, "ellipsis_mask", ellipsis_mask,
"new_axis_mask", new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
return _result
except _core._FallbackException:
try:
return strided_slice_grad_eager_fallback(
shape, begin, end, strides, dy, begin_mask=begin_mask,
end_mask=end_mask, ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask, shrink_axis_mask=shrink_axis_mask,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"StridedSliceGrad", shape=shape, begin=begin, end=end,
strides=strides, dy=dy, begin_mask=begin_mask,
end_mask=end_mask, ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask,
shrink_axis_mask=shrink_axis_mask, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Index",
_op._get_attr_type("Index"), "begin_mask",
_op._get_attr_int("begin_mask"), "end_mask",
_op._get_attr_int("end_mask"), "ellipsis_mask",
_op._get_attr_int("ellipsis_mask"), "new_axis_mask",
_op._get_attr_int("new_axis_mask"), "shrink_axis_mask",
_op._get_attr_int("shrink_axis_mask"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"StridedSliceGrad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
StridedSliceGrad = tf_export("raw_ops.StridedSliceGrad")(_ops.to_raw_op(strided_slice_grad))
def strided_slice_grad_eager_fallback(shape, begin, end, strides, dy, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask, name, ctx):
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_attr_T, (dy,) = _execute.args_to_matching_eager([dy], ctx)
_attr_Index, _inputs_Index = _execute.args_to_matching_eager([shape, begin, end, strides], ctx)
(shape, begin, end, strides) = _inputs_Index
_inputs_flat = [shape, begin, end, strides, dy]
_attrs = ("T", _attr_T, "Index", _attr_Index, "begin_mask", begin_mask,
"end_mask", end_mask, "ellipsis_mask", ellipsis_mask, "new_axis_mask",
new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
_result = _execute.execute(b"StridedSliceGrad", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"StridedSliceGrad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('tensor_scatter_nd_add', v1=['tensor_scatter_nd_add', 'tensor_scatter_add'])
@deprecated_endpoints('tensor_scatter_add')
def tensor_scatter_add(tensor, indices, updates, name=None):
r"""Adds sparse `updates` to an existing tensor according to `indices`.
This operation creates a new tensor by adding sparse `updates` to the passed
in `tensor`.
This operation is very similar to `tf.scatter_nd_add`, except that the updates
are added onto an existing tensor (as opposed to a variable). If the memory
for the existing tensor cannot be re-used, a copy is made and updated.
`indices` is an integer tensor containing indices into a new tensor of shape
`shape`. The last dimension of `indices` can be at most the rank of `shape`:
indices.shape[-1] <= shape.rank
The last dimension of `indices` corresponds to indices into elements
(if `indices.shape[-1] = shape.rank`) or slices
(if `indices.shape[-1] < shape.rank`) along dimension `indices.shape[-1]` of
`shape`. `updates` is a tensor with shape
indices.shape[:-1] + shape[indices.shape[-1]:]
The simplest form of tensor_scatter_add is to add individual elements to a
tensor by index. For example, say we want to add 4 elements in a rank-1
tensor with 8 elements.
In Python, this scatter add operation would look like this:
```python
indices = tf.constant([[4], [3], [1], [7]])
updates = tf.constant([9, 10, 11, 12])
tensor = tf.ones([8], dtype=tf.int32)
updated = tf.tensor_scatter_nd_add(tensor, indices, updates)
print(updated)
```
The resulting tensor would look like this:
[1, 12, 1, 11, 10, 1, 1, 13]
We can also, insert entire slices of a higher rank tensor all at once. For
example, if we wanted to insert two slices in the first dimension of a
rank-3 tensor with two matrices of new values.
In Python, this scatter add operation would look like this:
```python
indices = tf.constant([[0], [2]])
updates = tf.constant([[[5, 5, 5, 5], [6, 6, 6, 6],
[7, 7, 7, 7], [8, 8, 8, 8]],
[[5, 5, 5, 5], [6, 6, 6, 6],
[7, 7, 7, 7], [8, 8, 8, 8]]])
tensor = tf.ones([4, 4, 4],dtype=tf.int32)
updated = tf.tensor_scatter_nd_add(tensor, indices, updates)
print(updated)
```
The resulting tensor would look like this:
[[[6, 6, 6, 6], [7, 7, 7, 7], [8, 8, 8, 8], [9, 9, 9, 9]],
[[1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1]],
[[6, 6, 6, 6], [7, 7, 7, 7], [8, 8, 8, 8], [9, 9, 9, 9]],
[[1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1]]]
Note that on CPU, if an out of bound index is found, an error is returned.
On GPU, if an out of bound index is found, the index is ignored.
Args:
tensor: A `Tensor`. Tensor to copy/update.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Index tensor.
updates: A `Tensor`. Must have the same type as `tensor`.
Updates to scatter into output.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "TensorScatterAdd", name,
tld.op_callbacks, tensor, indices, updates)
return _result
except _core._FallbackException:
try:
return tensor_scatter_add_eager_fallback(
tensor, indices, updates, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
tensor_scatter_add, tensor=tensor, indices=indices,
updates=updates, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"TensorScatterAdd", tensor=tensor, indices=indices, updates=updates,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
tensor_scatter_add, tensor=tensor, indices=indices, updates=updates,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"TensorScatterAdd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
TensorScatterAdd = tf_export("raw_ops.TensorScatterAdd")(_ops.to_raw_op(tensor_scatter_add))
def tensor_scatter_add_eager_fallback(tensor, indices, updates, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([tensor, updates], ctx)
(tensor, updates) = _inputs_T
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_inputs_flat = [tensor, indices, updates]
_attrs = ("T", _attr_T, "Tindices", _attr_Tindices)
_result = _execute.execute(b"TensorScatterAdd", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"TensorScatterAdd", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('tensor_scatter_nd_sub', v1=['tensor_scatter_nd_sub', 'tensor_scatter_sub'])
@deprecated_endpoints('tensor_scatter_sub')
def tensor_scatter_sub(tensor, indices, updates, name=None):
r"""Subtracts sparse `updates` from an existing tensor according to `indices`.
This operation creates a new tensor by subtracting sparse `updates` from the
passed in `tensor`.
This operation is very similar to `tf.scatter_nd_sub`, except that the updates
are subtracted from an existing tensor (as opposed to a variable). If the memory
for the existing tensor cannot be re-used, a copy is made and updated.
`indices` is an integer tensor containing indices into a new tensor of shape
`shape`. The last dimension of `indices` can be at most the rank of `shape`:
indices.shape[-1] <= shape.rank
The last dimension of `indices` corresponds to indices into elements
(if `indices.shape[-1] = shape.rank`) or slices
(if `indices.shape[-1] < shape.rank`) along dimension `indices.shape[-1]` of
`shape`. `updates` is a tensor with shape
indices.shape[:-1] + shape[indices.shape[-1]:]
The simplest form of tensor_scatter_sub is to subtract individual elements
from a tensor by index. For example, say we want to insert 4 scattered elements
in a rank-1 tensor with 8 elements.
In Python, this scatter subtract operation would look like this:
```python
indices = tf.constant([[4], [3], [1], [7]])
updates = tf.constant([9, 10, 11, 12])
tensor = tf.ones([8], dtype=tf.int32)
updated = tf.tensor_scatter_nd_sub(tensor, indices, updates)
print(updated)
```
The resulting tensor would look like this:
[1, -10, 1, -9, -8, 1, 1, -11]
We can also, insert entire slices of a higher rank tensor all at once. For
example, if we wanted to insert two slices in the first dimension of a
rank-3 tensor with two matrices of new values.
In Python, this scatter add operation would look like this:
```python
indices = tf.constant([[0], [2]])
updates = tf.constant([[[5, 5, 5, 5], [6, 6, 6, 6],
[7, 7, 7, 7], [8, 8, 8, 8]],
[[5, 5, 5, 5], [6, 6, 6, 6],
[7, 7, 7, 7], [8, 8, 8, 8]]])
tensor = tf.ones([4, 4, 4],dtype=tf.int32)
updated = tf.tensor_scatter_nd_sub(tensor, indices, updates)
print(updated)
```
The resulting tensor would look like this:
[[[-4, -4, -4, -4], [-5, -5, -5, -5], [-6, -6, -6, -6], [-7, -7, -7, -7]],
[[1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1]],
[[-4, -4, -4, -4], [-5, -5, -5, -5], [-6, -6, -6, -6], [-7, -7, -7, -7]],
[[1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1], [1, 1, 1, 1]]]
Note that on CPU, if an out of bound index is found, an error is returned.
On GPU, if an out of bound index is found, the index is ignored.
Args:
tensor: A `Tensor`. Tensor to copy/update.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Index tensor.
updates: A `Tensor`. Must have the same type as `tensor`.
Updates to scatter into output.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "TensorScatterSub", name,
tld.op_callbacks, tensor, indices, updates)
return _result
except _core._FallbackException:
try:
return tensor_scatter_sub_eager_fallback(
tensor, indices, updates, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
tensor_scatter_sub, tensor=tensor, indices=indices,
updates=updates, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"TensorScatterSub", tensor=tensor, indices=indices, updates=updates,
name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
tensor_scatter_sub, tensor=tensor, indices=indices, updates=updates,
name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"TensorScatterSub", _inputs_flat, _attrs, _result)
_result, = _result
return _result
TensorScatterSub = tf_export("raw_ops.TensorScatterSub")(_ops.to_raw_op(tensor_scatter_sub))
def tensor_scatter_sub_eager_fallback(tensor, indices, updates, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([tensor, updates], ctx)
(tensor, updates) = _inputs_T
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_inputs_flat = [tensor, indices, updates]
_attrs = ("T", _attr_T, "Tindices", _attr_Tindices)
_result = _execute.execute(b"TensorScatterSub", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"TensorScatterSub", _inputs_flat, _attrs, _result)
_result, = _result
return _result
@_dispatch.add_dispatch_list
@tf_export('tensor_scatter_nd_update', v1=['tensor_scatter_nd_update', 'tensor_scatter_update'])
@deprecated_endpoints('tensor_scatter_update')
def tensor_scatter_update(tensor, indices, updates, name=None):
r"""Scatter `updates` into an existing tensor according to `indices`.
This operation creates a new tensor by applying sparse `updates` to the passed
in `tensor`.
This operation is very similar to `tf.scatter_nd`, except that the updates are
scattered onto an existing tensor (as opposed to a zero-tensor). If the memory
for the existing tensor cannot be re-used, a copy is made and updated.
If `indices` contains duplicates, then their updates are accumulated (summed).
**WARNING**: The order in which updates are applied is nondeterministic, so the
output will be nondeterministic if `indices` contains duplicates -- because
of some numerical approximation issues, numbers summed in different order
may yield different results.
`indices` is an integer tensor containing indices into a new tensor of shape
`shape`. The last dimension of `indices` can be at most the rank of `shape`:
indices.shape[-1] <= shape.rank
The last dimension of `indices` corresponds to indices into elements
(if `indices.shape[-1] = shape.rank`) or slices
(if `indices.shape[-1] < shape.rank`) along dimension `indices.shape[-1]` of
`shape`. `updates` is a tensor with shape
indices.shape[:-1] + shape[indices.shape[-1]:]
The simplest form of scatter is to insert individual elements in a tensor by
index. For example, say we want to insert 4 scattered elements in a rank-1
tensor with 8 elements.
<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="https://www.tensorflow.org/images/ScatterNd1.png" alt>
</div>
In Python, this scatter operation would look like this:
>>> indices = tf.constant([[4], [3], [1], [7]])
>>> updates = tf.constant([9, 10, 11, 12])
>>> tensor = tf.ones([8], dtype=tf.int32)
>>> print(tf.tensor_scatter_nd_update(tensor, indices, updates))
tf.Tensor([ 1 11 1 10 9 1 1 12], shape=(8,), dtype=int32)
We can also, insert entire slices of a higher rank tensor all at once. For
example, if we wanted to insert two slices in the first dimension of a
rank-3 tensor with two matrices of new values.
In Python, this scatter operation would look like this:
>>> indices = tf.constant([[0], [2]])
>>> updates = tf.constant([[[5, 5, 5, 5], [6, 6, 6, 6],
... [7, 7, 7, 7], [8, 8, 8, 8]],
... [[5, 5, 5, 5], [6, 6, 6, 6],
... [7, 7, 7, 7], [8, 8, 8, 8]]])
>>> tensor = tf.ones([4, 4, 4], dtype=tf.int32)
>>> print(tf.tensor_scatter_nd_update(tensor, indices, updates).numpy())
[[[5 5 5 5]
[6 6 6 6]
[7 7 7 7]
[8 8 8 8]]
[[1 1 1 1]
[1 1 1 1]
[1 1 1 1]
[1 1 1 1]]
[[5 5 5 5]
[6 6 6 6]
[7 7 7 7]
[8 8 8 8]]
[[1 1 1 1]
[1 1 1 1]
[1 1 1 1]
[1 1 1 1]]]
Note that on CPU, if an out of bound index is found, an error is returned.
On GPU, if an out of bound index is found, the index is ignored.
Args:
tensor: A `Tensor`. Tensor to copy/update.
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
Index tensor.
updates: A `Tensor`. Must have the same type as `tensor`.
Updates to scatter into output.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `tensor`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "TensorScatterUpdate", name,
tld.op_callbacks, tensor, indices, updates)
return _result
except _core._FallbackException:
try:
return tensor_scatter_update_eager_fallback(
tensor, indices, updates, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
tensor_scatter_update, tensor=tensor, indices=indices,
updates=updates, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"TensorScatterUpdate", tensor=tensor, indices=indices,
updates=updates, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
tensor_scatter_update, tensor=tensor, indices=indices,
updates=updates, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tindices",
_op._get_attr_type("Tindices"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"TensorScatterUpdate", _inputs_flat, _attrs, _result)
_result, = _result
return _result
TensorScatterUpdate = tf_export("raw_ops.TensorScatterUpdate")(_ops.to_raw_op(tensor_scatter_update))
def tensor_scatter_update_eager_fallback(tensor, indices, updates, name, ctx):
_attr_T, _inputs_T = _execute.args_to_matching_eager([tensor, updates], ctx)
(tensor, updates) = _inputs_T
_attr_Tindices, (indices,) = _execute.args_to_matching_eager([indices], ctx)
_inputs_flat = [tensor, indices, updates]
_attrs = ("T", _attr_T, "Tindices", _attr_Tindices)
_result = _execute.execute(b"TensorScatterUpdate", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"TensorScatterUpdate", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def tensor_strided_slice_update(input, begin, end, strides, value, begin_mask=0, end_mask=0, ellipsis_mask=0, new_axis_mask=0, shrink_axis_mask=0, name=None):
r"""Assign `value` to the sliced l-value reference of `input`.
The values of `value` are assigned to the positions in the tensor `input` that
are selected by the slice parameters. The slice parameters `begin` `end`
`strides` etc. work exactly as in `StridedSlice`.
NOTE this op currently does not support broadcasting and so `value`'s shape
must be exactly the shape produced by the slice of `input`.
Args:
input: A `Tensor`.
begin: A `Tensor`. Must be one of the following types: `int32`, `int64`.
end: A `Tensor`. Must have the same type as `begin`.
strides: A `Tensor`. Must have the same type as `begin`.
value: A `Tensor`. Must have the same type as `input`.
begin_mask: An optional `int`. Defaults to `0`.
end_mask: An optional `int`. Defaults to `0`.
ellipsis_mask: An optional `int`. Defaults to `0`.
new_axis_mask: An optional `int`. Defaults to `0`.
shrink_axis_mask: An optional `int`. Defaults to `0`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "TensorStridedSliceUpdate",
name, tld.op_callbacks, input, begin, end, strides, value,
"begin_mask", begin_mask, "end_mask", end_mask, "ellipsis_mask",
ellipsis_mask, "new_axis_mask", new_axis_mask, "shrink_axis_mask",
shrink_axis_mask)
return _result
except _core._FallbackException:
try:
return tensor_strided_slice_update_eager_fallback(
input, begin, end, strides, value, begin_mask=begin_mask,
end_mask=end_mask, ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask, shrink_axis_mask=shrink_axis_mask,
name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"TensorStridedSliceUpdate", input=input, begin=begin, end=end,
strides=strides, value=value,
begin_mask=begin_mask, end_mask=end_mask,
ellipsis_mask=ellipsis_mask,
new_axis_mask=new_axis_mask,
shrink_axis_mask=shrink_axis_mask,
name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Index",
_op._get_attr_type("Index"), "begin_mask",
_op._get_attr_int("begin_mask"), "end_mask",
_op._get_attr_int("end_mask"), "ellipsis_mask",
_op._get_attr_int("ellipsis_mask"), "new_axis_mask",
_op._get_attr_int("new_axis_mask"), "shrink_axis_mask",
_op._get_attr_int("shrink_axis_mask"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"TensorStridedSliceUpdate", _inputs_flat, _attrs, _result)
_result, = _result
return _result
TensorStridedSliceUpdate = tf_export("raw_ops.TensorStridedSliceUpdate")(_ops.to_raw_op(tensor_strided_slice_update))
def tensor_strided_slice_update_eager_fallback(input, begin, end, strides, value, begin_mask, end_mask, ellipsis_mask, new_axis_mask, shrink_axis_mask, name, ctx):
if begin_mask is None:
begin_mask = 0
begin_mask = _execute.make_int(begin_mask, "begin_mask")
if end_mask is None:
end_mask = 0
end_mask = _execute.make_int(end_mask, "end_mask")
if ellipsis_mask is None:
ellipsis_mask = 0
ellipsis_mask = _execute.make_int(ellipsis_mask, "ellipsis_mask")
if new_axis_mask is None:
new_axis_mask = 0
new_axis_mask = _execute.make_int(new_axis_mask, "new_axis_mask")
if shrink_axis_mask is None:
shrink_axis_mask = 0
shrink_axis_mask = _execute.make_int(shrink_axis_mask, "shrink_axis_mask")
_attr_T, _inputs_T = _execute.args_to_matching_eager([input, value], ctx)
(input, value) = _inputs_T
_attr_Index, _inputs_Index = _execute.args_to_matching_eager([begin, end, strides], ctx)
(begin, end, strides) = _inputs_Index
_inputs_flat = [input, begin, end, strides, value]
_attrs = ("T", _attr_T, "Index", _attr_Index, "begin_mask", begin_mask,
"end_mask", end_mask, "ellipsis_mask", ellipsis_mask, "new_axis_mask",
new_axis_mask, "shrink_axis_mask", shrink_axis_mask)
_result = _execute.execute(b"TensorStridedSliceUpdate", 1,
inputs=_inputs_flat, attrs=_attrs, ctx=ctx,
name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"TensorStridedSliceUpdate", _inputs_flat, _attrs, _result)
_result, = _result
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('tile', v1=['tile', 'manip.tile'])
@deprecated_endpoints('manip.tile')
def tile(input, multiples, name=None):
r"""Constructs a tensor by tiling a given tensor.
This operation creates a new tensor by replicating `input` `multiples` times.
The output tensor's i'th dimension has `input.dims(i) * multiples[i]` elements,
and the values of `input` are replicated `multiples[i]` times along the 'i'th
dimension. For example, tiling `[a b c d]` by `[2]` produces
`[a b c d a b c d]`.
>>> a = tf.constant([[1,2,3],[4,5,6]], tf.int32)
>>> b = tf.constant([1,2], tf.int32)
>>> tf.tile(a, b)
<tf.Tensor: shape=(2, 6), dtype=int32, numpy=
array([[1, 2, 3, 1, 2, 3],
[4, 5, 6, 4, 5, 6]], dtype=int32)>
>>> c = tf.constant([2,1], tf.int32)
>>> tf.tile(a, c)
<tf.Tensor: shape=(4, 3), dtype=int32, numpy=
array([[1, 2, 3],
[4, 5, 6],
[1, 2, 3],
[4, 5, 6]], dtype=int32)>
>>> d = tf.constant([2,2], tf.int32)
>>> tf.tile(a, d)
<tf.Tensor: shape=(4, 6), dtype=int32, numpy=
array([[1, 2, 3, 1, 2, 3],
[4, 5, 6, 4, 5, 6],
[1, 2, 3, 1, 2, 3],
[4, 5, 6, 4, 5, 6]], dtype=int32)>
Args:
input: A `Tensor`. 1-D or higher.
multiples: A `Tensor`. Must be one of the following types: `int32`, `int64`.
1-D. Length must be the same as the number of dimensions in `input`
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Tile", name, tld.op_callbacks,
input, multiples)
return _result
except _core._FallbackException:
try:
return tile_eager_fallback(
input, multiples, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
tile, input=input, multiples=multiples, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Tile", input=input, multiples=multiples, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
tile, input=input, multiples=multiples, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tmultiples",
_op._get_attr_type("Tmultiples"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Tile", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Tile = tf_export("raw_ops.Tile")(_ops.to_raw_op(tile))
def tile_eager_fallback(input, multiples, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
_attr_Tmultiples, (multiples,) = _execute.args_to_matching_eager([multiples], ctx, _dtypes.int32)
_inputs_flat = [input, multiples]
_attrs = ("T", _attr_T, "Tmultiples", _attr_Tmultiples)
_result = _execute.execute(b"Tile", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Tile", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def tile_grad(input, multiples, name=None):
r"""Returns the gradient of `Tile`.
Since `Tile` takes an input and repeats the input `multiples` times
along each dimension, `TileGrad` takes in `multiples` and aggregates
each repeated tile of `input` into `output`.
Args:
input: A `Tensor`.
multiples: A `Tensor` of type `int32`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `input`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "TileGrad", name,
tld.op_callbacks, input, multiples)
return _result
except _core._FallbackException:
try:
return tile_grad_eager_fallback(
input, multiples, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"TileGrad", input=input, multiples=multiples, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"TileGrad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
TileGrad = tf_export("raw_ops.TileGrad")(_ops.to_raw_op(tile_grad))
def tile_grad_eager_fallback(input, multiples, name, ctx):
_attr_T, (input,) = _execute.args_to_matching_eager([input], ctx)
multiples = _ops.convert_to_tensor(multiples, _dtypes.int32)
_inputs_flat = [input, multiples]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"TileGrad", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"TileGrad", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def transpose(x, perm, name=None):
r"""Shuffle dimensions of x according to a permutation.
The output `y` has the same rank as `x`. The shapes of `x` and `y` satisfy:
`y.shape[i] == x.shape[perm[i]] for i in [0, 1, ..., rank(x) - 1]`
Args:
x: A `Tensor`.
perm: A `Tensor`. Must be one of the following types: `int32`, `int64`.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Transpose", name,
tld.op_callbacks, x, perm)
return _result
except _core._FallbackException:
try:
return transpose_eager_fallback(
x, perm, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Transpose", x=x, perm=perm, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Tperm",
_op._get_attr_type("Tperm"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Transpose", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Transpose = tf_export("raw_ops.Transpose")(_ops.to_raw_op(transpose))
def transpose_eager_fallback(x, perm, name, ctx):
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_attr_Tperm, (perm,) = _execute.args_to_matching_eager([perm], ctx, _dtypes.int32)
_inputs_flat = [x, perm]
_attrs = ("T", _attr_T, "Tperm", _attr_Tperm)
_result = _execute.execute(b"Transpose", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Transpose", _inputs_flat, _attrs, _result)
_result, = _result
return _result
_UniqueOutput = collections.namedtuple(
"Unique",
["y", "idx"])
def unique(x, out_idx=_dtypes.int32, name=None):
r"""Finds unique elements in a 1-D tensor.
This operation returns a tensor `y` containing all of the unique elements of `x`
sorted in the same order that they occur in `x`; `x` does not need to be sorted.
This operation also returns a tensor `idx` the same size as `x` that contains
the index of each value of `x` in the unique output `y`. In other words:
`y[idx[i]] = x[i] for i in [0, 1,...,rank(x) - 1]`
Examples:
```
# tensor 'x' is [1, 1, 2, 4, 4, 4, 7, 8, 8]
y, idx = unique(x)
y ==> [1, 2, 4, 7, 8]
idx ==> [0, 0, 1, 2, 2, 2, 3, 4, 4]
```
```
# tensor 'x' is [4, 5, 1, 2, 3, 3, 4, 5]
y, idx = unique(x)
y ==> [4, 5, 1, 2, 3]
idx ==> [0, 1, 2, 3, 4, 4, 0, 1]
```
Args:
x: A `Tensor`. 1-D.
out_idx: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (y, idx).
y: A `Tensor`. Has the same type as `x`.
idx: A `Tensor` of type `out_idx`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Unique", name,
tld.op_callbacks, x, "out_idx", out_idx)
_result = _UniqueOutput._make(_result)
return _result
except _core._FallbackException:
try:
return unique_eager_fallback(
x, out_idx=out_idx, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Unique", x=x, out_idx=out_idx, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_idx",
_op._get_attr_type("out_idx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Unique", _inputs_flat, _attrs, _result)
_result = _UniqueOutput._make(_result)
return _result
Unique = tf_export("raw_ops.Unique")(_ops.to_raw_op(unique))
def unique_eager_fallback(x, out_idx, name, ctx):
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_inputs_flat = [x]
_attrs = ("T", _attr_T, "out_idx", out_idx)
_result = _execute.execute(b"Unique", 2, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Unique", _inputs_flat, _attrs, _result)
_result = _UniqueOutput._make(_result)
return _result
_UniqueV2Output = collections.namedtuple(
"UniqueV2",
["y", "idx"])
def unique_v2(x, axis, out_idx=_dtypes.int32, name=None):
r"""Finds unique elements along an axis of a tensor.
This operation either returns a tensor `y` containing unique elements
along the `axis` of a tensor. The returned unique elements is sorted
in the same order as they occur along `axis` in `x`.
This operation also returns a tensor `idx` that is the same size as
the number of the elements in `x` along the `axis` dimension. It
contains the index in the unique output `y`.
In other words, for an `1-D` tensor `x` with `axis = None:
`y[idx[i]] = x[i] for i in [0, 1,...,rank(x) - 1]`
For example:
```
# tensor 'x' is [1, 1, 2, 4, 4, 4, 7, 8, 8]
y, idx = unique(x)
y ==> [1, 2, 4, 7, 8]
idx ==> [0, 0, 1, 2, 2, 2, 3, 4, 4]
```
For an `2-D` tensor `x` with `axis = 0`:
```
# tensor 'x' is [[1, 0, 0],
# [1, 0, 0],
# [2, 0, 0]]
y, idx = unique(x, axis=0)
y ==> [[1, 0, 0],
[2, 0, 0]]
idx ==> [0, 0, 1]
```
For an `2-D` tensor `x` with `axis = 1`:
```
# tensor 'x' is [[1, 0, 0],
# [1, 0, 0],
# [2, 0, 0]]
y, idx = unique(x, axis=1)
y ==> [[1, 0],
[1, 0],
[2, 0]]
idx ==> [0, 1, 1]
```
Args:
x: A `Tensor`. A `Tensor`.
axis: A `Tensor`. Must be one of the following types: `int32`, `int64`.
A `Tensor` of type `int32` (default: None). The axis of the Tensor to
find the unique elements.
out_idx: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (y, idx).
y: A `Tensor`. Has the same type as `x`.
idx: A `Tensor` of type `out_idx`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "UniqueV2", name,
tld.op_callbacks, x, axis, "out_idx", out_idx)
_result = _UniqueV2Output._make(_result)
return _result
except _core._FallbackException:
try:
return unique_v2_eager_fallback(
x, axis, out_idx=out_idx, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"UniqueV2", x=x, axis=axis, out_idx=out_idx, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Taxis",
_op._get_attr_type("Taxis"), "out_idx",
_op._get_attr_type("out_idx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"UniqueV2", _inputs_flat, _attrs, _result)
_result = _UniqueV2Output._make(_result)
return _result
UniqueV2 = tf_export("raw_ops.UniqueV2")(_ops.to_raw_op(unique_v2))
def unique_v2_eager_fallback(x, axis, out_idx, name, ctx):
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_attr_Taxis, (axis,) = _execute.args_to_matching_eager([axis], ctx, _dtypes.int64)
_inputs_flat = [x, axis]
_attrs = ("T", _attr_T, "Taxis", _attr_Taxis, "out_idx", out_idx)
_result = _execute.execute(b"UniqueV2", 2, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"UniqueV2", _inputs_flat, _attrs, _result)
_result = _UniqueV2Output._make(_result)
return _result
_UniqueWithCountsOutput = collections.namedtuple(
"UniqueWithCounts",
["y", "idx", "count"])
def unique_with_counts(x, out_idx=_dtypes.int32, name=None):
r"""Finds unique elements in a 1-D tensor.
This operation returns a tensor `y` containing all of the unique elements of `x`
sorted in the same order that they occur in `x`. This operation also returns a
tensor `idx` the same size as `x` that contains the index of each value of `x`
in the unique output `y`. Finally, it returns a third tensor `count` that
contains the count of each element of `y` in `x`. In other words:
`y[idx[i]] = x[i] for i in [0, 1,...,rank(x) - 1]`
For example:
```
# tensor 'x' is [1, 1, 2, 4, 4, 4, 7, 8, 8]
y, idx, count = unique_with_counts(x)
y ==> [1, 2, 4, 7, 8]
idx ==> [0, 0, 1, 2, 2, 2, 3, 4, 4]
count ==> [2, 1, 3, 1, 2]
```
Args:
x: A `Tensor`. 1-D.
out_idx: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (y, idx, count).
y: A `Tensor`. Has the same type as `x`.
idx: A `Tensor` of type `out_idx`.
count: A `Tensor` of type `out_idx`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "UniqueWithCounts", name,
tld.op_callbacks, x, "out_idx", out_idx)
_result = _UniqueWithCountsOutput._make(_result)
return _result
except _core._FallbackException:
try:
return unique_with_counts_eager_fallback(
x, out_idx=out_idx, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"UniqueWithCounts", x=x, out_idx=out_idx, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_idx",
_op._get_attr_type("out_idx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"UniqueWithCounts", _inputs_flat, _attrs, _result)
_result = _UniqueWithCountsOutput._make(_result)
return _result
UniqueWithCounts = tf_export("raw_ops.UniqueWithCounts")(_ops.to_raw_op(unique_with_counts))
def unique_with_counts_eager_fallback(x, out_idx, name, ctx):
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_inputs_flat = [x]
_attrs = ("T", _attr_T, "out_idx", out_idx)
_result = _execute.execute(b"UniqueWithCounts", 3, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"UniqueWithCounts", _inputs_flat, _attrs, _result)
_result = _UniqueWithCountsOutput._make(_result)
return _result
_UniqueWithCountsV2Output = collections.namedtuple(
"UniqueWithCountsV2",
["y", "idx", "count"])
def unique_with_counts_v2(x, axis, out_idx=_dtypes.int32, name=None):
r"""Finds unique elements along an axis of a tensor.
This operation either returns a tensor `y` containing unique elements
along the `axis` of a tensor. The returned unique elements is sorted
in the same order as they occur along `axis` in `x`.
This operation also returns a tensor `idx` and a tensor `count`
that are the same size as the number of the elements in `x` along the
`axis` dimension. The `idx` contains the index in the unique output `y`
and the `count` contains the count in the unique output `y`.
In other words, for an `1-D` tensor `x` with `axis = None:
`y[idx[i]] = x[i] for i in [0, 1,...,rank(x) - 1]`
For example:
```
# tensor 'x' is [1, 1, 2, 4, 4, 4, 7, 8, 8]
y, idx, count = unique_with_counts(x)
y ==> [1, 2, 4, 7, 8]
idx ==> [0, 0, 1, 2, 2, 2, 3, 4, 4]
count ==> [2, 1, 3, 1, 2]
```
For an `2-D` tensor `x` with `axis = 0`:
```
# tensor 'x' is [[1, 0, 0],
# [1, 0, 0],
# [2, 0, 0]]
y, idx, count = unique_with_counts(x, axis=0)
y ==> [[1, 0, 0],
[2, 0, 0]]
idx ==> [0, 0, 1]
count ==> [2, 1]
```
For an `2-D` tensor `x` with `axis = 1`:
```
# tensor 'x' is [[1, 0, 0],
# [1, 0, 0],
# [2, 0, 0]]
y, idx, count = unique_with_counts(x, axis=1)
y ==> [[1, 0],
[1, 0],
[2, 0]]
idx ==> [0, 1, 1]
count ==> [1, 2]
```
Args:
x: A `Tensor`. A `Tensor`.
axis: A `Tensor`. Must be one of the following types: `int32`, `int64`.
A `Tensor` of type `int32` (default: None). The axis of the Tensor to
find the unique elements.
out_idx: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A tuple of `Tensor` objects (y, idx, count).
y: A `Tensor`. Has the same type as `x`.
idx: A `Tensor` of type `out_idx`.
count: A `Tensor` of type `out_idx`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "UniqueWithCountsV2", name,
tld.op_callbacks, x, axis, "out_idx", out_idx)
_result = _UniqueWithCountsV2Output._make(_result)
return _result
except _core._FallbackException:
try:
return unique_with_counts_v2_eager_fallback(
x, axis, out_idx=out_idx, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"UniqueWithCountsV2", x=x, axis=axis, out_idx=out_idx, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "Taxis",
_op._get_attr_type("Taxis"), "out_idx",
_op._get_attr_type("out_idx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"UniqueWithCountsV2", _inputs_flat, _attrs, _result)
_result = _UniqueWithCountsV2Output._make(_result)
return _result
UniqueWithCountsV2 = tf_export("raw_ops.UniqueWithCountsV2")(_ops.to_raw_op(unique_with_counts_v2))
def unique_with_counts_v2_eager_fallback(x, axis, out_idx, name, ctx):
if out_idx is None:
out_idx = _dtypes.int32
out_idx = _execute.make_type(out_idx, "out_idx")
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_attr_Taxis, (axis,) = _execute.args_to_matching_eager([axis], ctx, _dtypes.int64)
_inputs_flat = [x, axis]
_attrs = ("T", _attr_T, "Taxis", _attr_Taxis, "out_idx", out_idx)
_result = _execute.execute(b"UniqueWithCountsV2", 3, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"UniqueWithCountsV2", _inputs_flat, _attrs, _result)
_result = _UniqueWithCountsV2Output._make(_result)
return _result
def unpack(value, num, axis=0, name=None):
r"""Unpacks a given dimension of a rank-`R` tensor into `num` rank-`(R-1)` tensors.
Unpacks `num` tensors from `value` by chipping it along the `axis` dimension.
For example, given a tensor of shape `(A, B, C, D)`;
If `axis == 0` then the i'th tensor in `output` is the slice `value[i, :, :, :]`
and each tensor in `output` will have shape `(B, C, D)`. (Note that the
dimension unpacked along is gone, unlike `split`).
If `axis == 1` then the i'th tensor in `output` is the slice `value[:, i, :, :]`
and each tensor in `output` will have shape `(A, C, D)`.
Etc.
This is the opposite of `pack`.
Args:
value: A `Tensor`.
1-D or higher, with `axis` dimension size equal to `num`.
num: An `int` that is `>= 0`.
axis: An optional `int`. Defaults to `0`.
Dimension along which to unpack. Negative values wrap around, so the
valid range is `[-R, R)`.
name: A name for the operation (optional).
Returns:
A list of `num` `Tensor` objects with the same type as `value`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Unpack", name,
tld.op_callbacks, value, "num", num, "axis", axis)
return _result
except _core._FallbackException:
try:
return unpack_eager_fallback(
value, num=num, axis=axis, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
num = _execute.make_int(num, "num")
if axis is None:
axis = 0
axis = _execute.make_int(axis, "axis")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Unpack", value=value, num=num, axis=axis, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("num", _op._get_attr_int("num"), "T", _op._get_attr_type("T"),
"axis", _op._get_attr_int("axis"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Unpack", _inputs_flat, _attrs, _result)
return _result
Unpack = tf_export("raw_ops.Unpack")(_ops.to_raw_op(unpack))
def unpack_eager_fallback(value, num, axis, name, ctx):
num = _execute.make_int(num, "num")
if axis is None:
axis = 0
axis = _execute.make_int(axis, "axis")
_attr_T, (value,) = _execute.args_to_matching_eager([value], ctx)
_inputs_flat = [value]
_attrs = ("num", num, "T", _attr_T, "axis", axis)
_result = _execute.execute(b"Unpack", num, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Unpack", _inputs_flat, _attrs, _result)
return _result
[文档]@_dispatch.add_dispatch_list
@tf_export('unravel_index')
def unravel_index(indices, dims, name=None):
r"""Converts an array of flat indices into a tuple of coordinate arrays.
Example:
```
y = tf.unravel_index(indices=[2, 5, 7], dims=[3, 3])
# 'dims' represent a hypothetical (3, 3) tensor of indices:
# [[0, 1, *2*],
# [3, 4, *5*],
# [6, *7*, 8]]
# For each entry from 'indices', this operation returns
# its coordinates (marked with '*'), such as
# 2 ==> (0, 2)
# 5 ==> (1, 2)
# 7 ==> (2, 1)
y ==> [[0, 1, 2], [2, 2, 1]]
```
@compatibility(numpy)
Equivalent to np.unravel_index
@end_compatibility
Args:
indices: A `Tensor`. Must be one of the following types: `int32`, `int64`.
An 0-D or 1-D `int` Tensor whose elements are indices into the
flattened version of an array of dimensions dims.
dims: A `Tensor`. Must have the same type as `indices`.
An 1-D `int` Tensor. The shape of the array to use for unraveling
indices.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `indices`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "UnravelIndex", name,
tld.op_callbacks, indices, dims)
return _result
except _core._FallbackException:
try:
return unravel_index_eager_fallback(
indices, dims, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except (TypeError, ValueError):
result = _dispatch.dispatch(
unravel_index, indices=indices, dims=dims, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
try:
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"UnravelIndex", indices=indices, dims=dims, name=name)
except (TypeError, ValueError):
result = _dispatch.dispatch(
unravel_index, indices=indices, dims=dims, name=name)
if result is not _dispatch.OpDispatcher.NOT_SUPPORTED:
return result
raise
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("Tidx", _op._get_attr_type("Tidx"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"UnravelIndex", _inputs_flat, _attrs, _result)
_result, = _result
return _result
UnravelIndex = tf_export("raw_ops.UnravelIndex")(_ops.to_raw_op(unravel_index))
def unravel_index_eager_fallback(indices, dims, name, ctx):
_attr_Tidx, _inputs_Tidx = _execute.args_to_matching_eager([indices, dims], ctx, _dtypes.int32)
(indices, dims) = _inputs_Tidx
_inputs_flat = [indices, dims]
_attrs = ("Tidx", _attr_Tidx)
_result = _execute.execute(b"UnravelIndex", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"UnravelIndex", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def upper_bound(sorted_inputs, values, out_type=_dtypes.int32, name=None):
r"""Applies upper_bound(sorted_search_values, values) along each row.
Each set of rows with the same index in (sorted_inputs, values) is treated
independently. The resulting row is the equivalent of calling
`np.searchsorted(sorted_inputs, values, side='right')`.
The result is not a global index to the entire
`Tensor`, but rather just the index in the last dimension.
A 2-D example:
sorted_sequence = [[0, 3, 9, 9, 10],
[1, 2, 3, 4, 5]]
values = [[2, 4, 9],
[0, 2, 6]]
result = UpperBound(sorted_sequence, values)
result == [[1, 2, 4],
[0, 2, 5]]
Args:
sorted_inputs: A `Tensor`. 2-D Tensor where each row is ordered.
values: A `Tensor`. Must have the same type as `sorted_inputs`.
2-D Tensor with the same numbers of rows as `sorted_search_values`. Contains
the values that will be searched for in `sorted_search_values`.
out_type: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `out_type`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "UpperBound", name,
tld.op_callbacks, sorted_inputs, values, "out_type", out_type)
return _result
except _core._FallbackException:
try:
return upper_bound_eager_fallback(
sorted_inputs, values, out_type=out_type, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"UpperBound", sorted_inputs=sorted_inputs, values=values,
out_type=out_type, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"), "out_type",
_op._get_attr_type("out_type"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"UpperBound", _inputs_flat, _attrs, _result)
_result, = _result
return _result
UpperBound = tf_export("raw_ops.UpperBound")(_ops.to_raw_op(upper_bound))
def upper_bound_eager_fallback(sorted_inputs, values, out_type, name, ctx):
if out_type is None:
out_type = _dtypes.int32
out_type = _execute.make_type(out_type, "out_type")
_attr_T, _inputs_T = _execute.args_to_matching_eager([sorted_inputs, values], ctx)
(sorted_inputs, values) = _inputs_T
_inputs_flat = [sorted_inputs, values]
_attrs = ("T", _attr_T, "out_type", out_type)
_result = _execute.execute(b"UpperBound", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"UpperBound", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def where(condition, name=None):
r"""Returns locations of nonzero / true values in a tensor.
This operation returns the coordinates of true elements in `condition`. The
coordinates are returned in a 2-D tensor where the first dimension (rows)
represents the number of true elements, and the second dimension (columns)
represents the coordinates of the true elements. Keep in mind, the shape of
the output tensor can vary depending on how many true values there are in
`condition`. Indices are output in row-major order.
For example:
```
# 'input' tensor is [[True, False]
# [True, False]]
# 'input' has two true values, so output has two coordinates.
# 'input' has rank of 2, so coordinates have two indices.
where(input) ==> [[0, 0],
[1, 0]]
# `condition` tensor is [[[True, False]
# [True, False]]
# [[False, True]
# [False, True]]
# [[False, False]
# [False, True]]]
# 'input' has 5 true values, so output has 5 coordinates.
# 'input' has rank of 3, so coordinates have three indices.
where(input) ==> [[0, 0, 0],
[0, 1, 0],
[1, 0, 1],
[1, 1, 1],
[2, 1, 1]]
# `condition` tensor is [[[1.5, 0.0]
# [-0.5, 0.0]]
# [[0.0, 0.25]
# [0.0, 0.75]]
# [[0.0, 0.0]
# [0.0, 0.01]]]
# 'input' has 5 nonzero values, so output has 5 coordinates.
# 'input' has rank of 3, so coordinates have three indices.
where(input) ==> [[0, 0, 0],
[0, 1, 0],
[1, 0, 1],
[1, 1, 1],
[2, 1, 1]]
# `condition` tensor is [[[1.5 + 0.0j, 0.0 + 0.0j]
# [0.0 + 0.5j, 0.0 + 0.0j]]
# [[0.0 + 0.0j, 0.25 + 1.5j]
# [0.0 + 0.0j, 0.75 + 0.0j]]
# [[0.0 + 0.0j, 0.0 + 0.0j]
# [0.0 + 0.0j, 0.01 + 0.0j]]]
# 'input' has 5 nonzero magnitude values, so output has 5 coordinates.
# 'input' has rank of 3, so coordinates have three indices.
where(input) ==> [[0, 0, 0],
[0, 1, 0],
[1, 0, 1],
[1, 1, 1],
[2, 1, 1]]
```
Args:
condition: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `uint8`, `int16`, `int8`, `complex64`, `int64`, `qint8`, `quint8`, `qint32`, `bfloat16`, `uint16`, `complex128`, `half`, `uint32`, `uint64`, `bool`.
name: A name for the operation (optional).
Returns:
A `Tensor` of type `int64`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "Where", name,
tld.op_callbacks, condition)
return _result
except _core._FallbackException:
try:
return where_eager_fallback(
condition, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"Where", input=condition, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"Where", _inputs_flat, _attrs, _result)
_result, = _result
return _result
Where = tf_export("raw_ops.Where")(_ops.to_raw_op(where))
def where_eager_fallback(condition, name, ctx):
_attr_T, (condition,) = _execute.args_to_matching_eager([condition], ctx, _dtypes.bool)
_inputs_flat = [condition]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"Where", 1, inputs=_inputs_flat, attrs=_attrs,
ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"Where", _inputs_flat, _attrs, _result)
_result, = _result
return _result
def zeros_like(x, name=None):
r"""Returns a tensor of zeros with the same shape and type as x.
Args:
x: A `Tensor`. a tensor of type T.
name: A name for the operation (optional).
Returns:
A `Tensor`. Has the same type as `x`.
"""
_ctx = _context._context or _context.context()
tld = _ctx._thread_local_data
if tld.is_eager:
try:
_result = pywrap_tfe.TFE_Py_FastPathExecute(
_ctx._context_handle, tld.device_name, "ZerosLike", name,
tld.op_callbacks, x)
return _result
except _core._FallbackException:
try:
return zeros_like_eager_fallback(
x, name=name, ctx=_ctx)
except _core._SymbolicException:
pass # Add nodes to the TensorFlow graph.
except _core._NotOkStatusException as e:
_ops.raise_from_not_ok_status(e, name)
# Add nodes to the TensorFlow graph.
_, _, _op, _outputs = _op_def_library._apply_op_helper(
"ZerosLike", x=x, name=name)
_result = _outputs[:]
if _execute.must_record_gradient():
_attrs = ("T", _op._get_attr_type("T"))
_inputs_flat = _op.inputs
_execute.record_gradient(
"ZerosLike", _inputs_flat, _attrs, _result)
_result, = _result
return _result
ZerosLike = tf_export("raw_ops.ZerosLike")(_ops.to_raw_op(zeros_like))
def zeros_like_eager_fallback(x, name, ctx):
_attr_T, (x,) = _execute.args_to_matching_eager([x], ctx)
_inputs_flat = [x]
_attrs = ("T", _attr_T)
_result = _execute.execute(b"ZerosLike", 1, inputs=_inputs_flat,
attrs=_attrs, ctx=ctx, name=name)
if _execute.must_record_gradient():
_execute.record_gradient(
"ZerosLike", _inputs_flat, _attrs, _result)
_result, = _result
return _result