# Math

Note: Functions taking `Tensor`

arguments can also take anything accepted by
`tf.convert_to_tensor`

.

[TOC]

## Arithmetic Operators

TensorFlow provides several operations that you can use to add basic arithmetic operators to your graph.

`tf.add(x, y, name=None)`

Returns x + y element-wise.

*NOTE*: Add supports broadcasting. AddN does not.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`uint8`

,`int8`

,`int16`

,`int32`

,`int64`

,`complex64`

,`complex128`

,`string`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.sub(x, y, name=None)`

Returns x - y element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.mul(x, y, name=None)`

Returns x * y element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`uint8`

,`int8`

,`int16`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.div(x, y, name=None)`

Returns x / y element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`uint8`

,`int8`

,`int16`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.truediv(x, y, name=None)`

Divides x / y elementwise, always producing floating point results.

The same as `tf.div`

for floating point arguments, but casts integer arguments
to floating point before dividing so that the result is always floating point.
This op is generated by normal `x / y`

division in Python 3 and in Python 2.7
with `from __future__ import division`

. If you want integer division that
rounds down, use `x // y`

or `tf.floordiv`

.

`x`

and `y`

must have the same numeric type. If the inputs are floating
point, the output will have the same type. If the inputs are integral, the
inputs are cast to `float32`

for `int8`

and `int16`

and `float64`

for `int32`

and `int64`

(matching the behavior of Numpy).

##### Args:

:`x`

`Tensor`

numerator of numeric type.:`y`

`Tensor`

denominator of numeric type.: A name for the operation (optional).`name`

##### Returns:

`x / y`

evaluated in floating point.

##### Raises:

: If`TypeError`

`x`

and`y`

have different dtypes.

`tf.floordiv(x, y, name=None)`

Divides `x / y`

elementwise, rounding down for floating point.

The same as `tf.div(x,y)`

for integers, but uses `tf.floor(tf.div(x,y))`

for
floating point arguments so that the result is always an integer (though
possibly an integer represented as floating point). This op is generated by
`x // y`

floor division in Python 3 and in Python 2.7 with
`from __future__ import division`

.

Note that for efficiency, `floordiv`

uses C semantics for negative numbers
(unlike Python and Numpy).

`x`

and `y`

must have the same type, and the result will have the same type
as well.

##### Args:

:`x`

`Tensor`

numerator of real numeric type.:`y`

`Tensor`

denominator of real numeric type.: A name for the operation (optional).`name`

##### Returns:

`x / y`

rounded down (except possibly towards zero for negative integers).

##### Raises:

: If the inputs are complex.`TypeError`

`tf.mod(x, y, name=None)`

Returns element-wise remainder of division.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

,`float32`

,`float64`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.cross(a, b, name=None)`

Compute the pairwise cross product.

`a`

and `b`

must be the same shape; they can either be simple 3-element vectors,
or any shape where the innermost dimension is 3. In the latter case, each pair
of corresponding 3-element vectors is cross-multiplied independently.

##### Args:

: A`a`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`uint8`

,`int16`

,`int8`

,`uint16`

,`half`

. A tensor containing 3-element vectors.: A`b`

`Tensor`

. Must have the same type as`a`

. Another tensor, of same type and shape as`a`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `a`

.
Pairwise cross product of the vectors in `a`

and `b`

.

## Basic Math Functions

TensorFlow provides several operations that you can use to add basic mathematical functions to your graph.

`tf.add_n(inputs, name=None)`

Add all input tensors element wise.

##### Args:

: A list of at least 1`inputs`

`Tensor`

objects of the same type in:`float32`

,`float64`

,`int64`

,`int32`

,`uint8`

,`uint16`

,`int16`

,`int8`

,`complex64`

,`complex128`

,`qint8`

,`quint8`

,`qint32`

,`half`

. Must all be the same size and shape.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `inputs`

.

`tf.abs(x, name=None)`

Computes the absolute value of a tensor.

Given a tensor of real numbers `x`

, this operation returns a tensor
containing the absolute value of each element in `x`

. For example, if x is
an input element and y is an output element, this operation computes
\(y = |x|\).

See `tf.complex_abs()`

to compute the absolute value of a complex
number.

##### Args:

: A`x`

`Tensor`

of type`float`

,`double`

,`int32`

, or`int64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

the same size and type as `x`

with absolute values.

`tf.neg(x, name=None)`

Computes numerical negative value element-wise.

I.e., \(y = -x\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.sign(x, name=None)`

Returns an element-wise indication of the sign of a number.

`y = sign(x) = -1`

if `x < 0`

; 0 if `x == 0`

; 1 if `x > 0`

.

For complex numbers, `y = sign(x) = x / |x|`

if `x != 0`

, otherwise `y = 0`

.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.inv(x, name=None)`

Computes the reciprocal of x element-wise.

I.e., \(y = 1 / x\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.square(x, name=None)`

Computes square of x element-wise.

I.e., \(y = x * x = x^2\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.round(x, name=None)`

Rounds the values of a tensor to the nearest integer, element-wise.

For example:

```
# 'a' is [0.9, 2.5, 2.3, -4.4]
tf.round(a) ==> [ 1.0, 3.0, 2.0, -4.0 ]
```

##### Args:

: A`x`

`Tensor`

of type`float`

or`double`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of same shape and type as `x`

.

`tf.sqrt(x, name=None)`

Computes square root of x element-wise.

I.e., \(y = \sqrt{x} = x^{1/2}\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.rsqrt(x, name=None)`

Computes reciprocal of square root of x element-wise.

I.e., \(y = 1 / \sqrt{x}\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.pow(x, y, name=None)`

Computes the power of one value to another.

Given a tensor `x`

and a tensor `y`

, this operation computes \(x^y\) for
corresponding elements in `x`

and `y`

. For example:

```
# tensor 'x' is [[2, 2], [3, 3]]
# tensor 'y' is [[8, 16], [2, 3]]
tf.pow(x, y) ==> [[256, 65536], [9, 27]]
```

##### Args:

: A`x`

`Tensor`

of type`float`

,`double`

,`int32`

,`int64`

,`complex64`

, or`complex128`

.: A`y`

`Tensor`

of type`float`

,`double`

,`int32`

,`int64`

,`complex64`

, or`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

.

`tf.exp(x, name=None)`

Computes exponential of x element-wise. \(y = e^x\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.log(x, name=None)`

Computes natural logarithm of x element-wise.

I.e., \(y = \log_e x\).

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.ceil(x, name=None)`

Returns element-wise smallest integer in not less than x.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.floor(x, name=None)`

Returns element-wise largest integer not greater than x.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.maximum(x, y, name=None)`

Returns the max of x and y (i.e. x > y ? x : y) element-wise, broadcasts.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.minimum(x, y, name=None)`

Returns the min of x and y (i.e. x < y ? x : y) element-wise, broadcasts.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.cos(x, name=None)`

Computes cos of x element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.sin(x, name=None)`

Computes sin of x element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.lbeta(x, name='lbeta')`

Computes `ln(|Beta(x)|)`

, reducing along the last dimension.

Given one-dimensional `z = [z_0,...,z_{K-1}]`

, we define

`Beta(z) = \prod_j Gamma(z_j) / Gamma(\sum_j z_j)`

And for `n + 1`

dimensional `x`

with shape `[N1, ..., Nn, K]`

, we define
`lbeta(x)[i1, ..., in] = Log(|Beta(x[i1, ..., in, :])|)`

. In other words,
the last dimension is treated as the `z`

vector.

Note that if `z = [u, v]`

, then
`Beta(z) = int_0^1 t^{u-1} (1 - t)^{v-1} dt`

, which defines the traditional
bivariate beta function.

##### Args:

: A rank`x`

`n + 1`

`Tensor`

with type`float`

, or`double`

.: A name for the operation (optional).`name`

##### Returns:

The logarithm of `|Beta(x)|`

reducing along the last dimension.

##### Raises:

: If`ValueError`

`x`

is empty with rank one or less.

`tf.tan(x, name=None)`

Computes tan of x element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.acos(x, name=None)`

Computes acos of x element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.asin(x, name=None)`

Computes asin of x element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.atan(x, name=None)`

Computes atan of x element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.lgamma(x, name=None)`

Computes the log of the absolute value of `Gamma(x)`

element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.digamma(x, name=None)`

Computes Psi, the derivative of Lgamma (the log of the absolute value of

`Gamma(x)`

), element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.erf(x, name=None)`

Computes the Gauss error function of `x`

element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.erfc(x, name=None)`

Computes the complementary error function of `x`

element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.squared_difference(x, y, name=None)`

Returns (x - y)(x - y) element-wise.

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

,`complex128`

.: A`y`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.igamma(a, x, name=None)`

Compute the lower regularized incomplete Gamma function `Q(a, x)`

.

The lower regularized incomplete Gamma function is defined as:

```
P(a, x) = gamma(a, x) / Gamma(x) = 1 - Q(a, x)
```

where

```
gamma(a, x) = int_{0}^{x} t^{a-1} exp(-t) dt
```

is the lower incomplete Gamma function.

Note, above `Q(a, x)`

(`Igammac`

) is the upper regularized complete
Gamma function.

##### Args:

: A`a`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`x`

`Tensor`

. Must have the same type as`a`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `a`

.

`tf.igammac(a, x, name=None)`

Compute the upper regularized incomplete Gamma function `Q(a, x)`

.

The upper regularized incomplete Gamma function is defined as:

```
Q(a, x) = Gamma(a, x) / Gamma(x) = 1 - P(a, x)
```

where

```
Gamma(a, x) = int_{x}^{\infty} t^{a-1} exp(-t) dt
```

is the upper incomplete Gama function.

Note, above `P(a, x)`

(`Igamma`

) is the lower regularized complete
Gamma function.

##### Args:

: A`a`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`x`

`Tensor`

. Must have the same type as`a`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `a`

.

`tf.zeta(x, q, name=None)`

Compute the Hurwitz zeta function \(\zeta(x, q)\).

The Hurwitz zeta function is defined as:

```
\zeta(x, q) = \sum_{n=0}^{\infty} (q + n)^{-x}
```

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`q`

`Tensor`

. Must have the same type as`x`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.

`tf.polygamma(a, x, name=None)`

Compute the polygamma function \(\psi^{(n)}(x)\).

The polygamma function is defined as:

```
\psi^{(n)}(x) = \frac{d^n}{dx^n} \psi(x)
```

where \(\psi(x)\) is the digamma function.

##### Args:

: A`a`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`x`

`Tensor`

. Must have the same type as`a`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `a`

.

## Matrix Math Functions

TensorFlow provides several operations that you can use to add linear algebra functions on matrices to your graph.

`tf.batch_matrix_diag(diagonal, name=None)`

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.batch_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:

: A`diagonal`

`Tensor`

. Rank`k`

, where`k >= 1`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `diagonal`

.
Rank `k+1`

, with `output.shape = diagonal.shape + [diagonal.shape[-1]]`

.

`tf.batch_matrix_diag_part(input, name=None)`

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, ..., N, N]`

, then the output is a
tensor of rank `k - 1`

with dimensions `[I, J, K, ..., 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.batch_matrix_diag_part(input) ==> [[1, 2, 3, 4], [5, 6, 7, 8]]
which has shape (2, 4)
```

##### Args:

: A`input`

`Tensor`

. Rank`k`

tensor where`k >= 2`

and the last two dimensions are equal.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

.
The extracted diagonal(s) having shape
`diagonal.shape = input.shape[:-1]`

.

`tf.batch_matrix_band_part(input, num_lower, num_upper, name=None)`

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)`is one if`

(num_lower < 0 || (m-n) <= num_lower)) &&
(num_upper < 0 || (n-m) <= num_upper)`, and zero otherwise.

For example:

```
# if 'input' is [[ 0, 1, 2, 3]
[-1, 0, 1, 2]
[-2, -1, 0, 1]
[-3, -2, -1, 0]],
tf.batch_matrix_band_part(input, 1, -1) ==> [[ 0, 1, 2, 3]
[-1, 0, 1, 2]
[ 0, -1, 0, 1]
[ 0, 0, -1, 0]],
tf.batch_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.batch_matrix_band_part(input, 0, -1) ==> Upper triangular part.
tf.batch_matrix_band_part(input, -1, 0) ==> Lower triangular part.
tf.batch_matrix_band_part(input, 0, 0) ==> Diagonal.
```

##### Args:

: A`input`

`Tensor`

. Rank`k`

tensor.: A`num_lower`

`Tensor`

of type`int64`

. 0-D tensor. Number of subdiagonals to keep. If negative, keep entire lower triangle.: A`num_upper`

`Tensor`

of type`int64`

. 0-D tensor. Number of superdiagonals to keep. If negative, keep entire upper triangle.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

.
Rank `k`

tensor of the same shape as input. The extracted banded tensor.

`tf.diag(diagonal, name=None)`

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:

: A`diagonal`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

. Rank k tensor where k is at most 3.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `diagonal`

.

`tf.diag_part(input, name=None)`

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:

: A`input`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`complex64`

. Rank k tensor where k is 2, 4, or 6.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. The extracted diagonal.

`tf.trace(x, name=None)`

Compute the trace of a tensor `x`

.

`trace(x)`

returns the sum of along the diagonal.

For example:

```
# 'x' is [[1, 1],
# [1, 1]]
tf.trace(x) ==> 2
# 'x' is [[1,2,3],
# [4,5,6],
# [7,8,9]]
tf.trace(x) ==> 15
```

##### Args:

: 2-D tensor.`x`

: A name for the operation (optional).`name`

##### Returns:

The trace of input tensor.

`tf.transpose(a, perm=None, name='transpose')`

Transposes `a`

. Permutes the dimensions according to `perm`

.

The returned tensor's dimension i will correspond to the input dimension
`perm[i]`

. If `perm`

is not given, it is set to (n-1...0), where n is
the rank of the input tensor. Hence by default, this operation performs a
regular matrix transpose on 2-D input Tensors.

For example:

```
# 'x' is [[1 2 3]
# [4 5 6]]
tf.transpose(x) ==> [[1 4]
[2 5]
[3 6]]
# Equivalently
tf.transpose(x, perm=[1, 0]) ==> [[1 4]
[2 5]
[3 6]]
# 'perm' is more useful for n-dimensional tensors, for n > 2
# 'x' is [[[1 2 3]
# [4 5 6]]
# [[7 8 9]
# [10 11 12]]]
# Take the transpose of the matrices in dimension-0
tf.transpose(x, perm=[0, 2, 1]) ==> [[[1 4]
[2 5]
[3 6]]
[[7 10]
[8 11]
[9 12]]]
```

##### Args:

: A`a`

`Tensor`

.: A permutation of the dimensions of`perm`

`a`

.: A name for the operation (optional).`name`

##### Returns:

A transposed `Tensor`

.

`tf.matmul(a, b, transpose_a=False, transpose_b=False, a_is_sparse=False, b_is_sparse=False, name=None)`

Multiplies matrix `a`

by matrix `b`

, producing `a`

* `b`

.

The inputs must be two-dimensional matrices, with matching inner dimensions, possibly after transposition.

Both matrices must be of the same type. The supported types are:
`float`

, `double`

, `int32`

, `complex64`

.

Either matrix can be transposed on the fly by setting the corresponding flag
to `True`

. This is `False`

by default.

If one or both of the matrices contain a lot of zeros, a more efficient
multiplication algorithm can be used by setting the corresponding
`a_is_sparse`

or `b_is_sparse`

flag to `True`

. These are `False`

by default.

For example:

```
# 2-D tensor `a`
a = tf.constant([1, 2, 3, 4, 5, 6], shape=[2, 3]) => [[1. 2. 3.]
[4. 5. 6.]]
# 2-D tensor `b`
b = tf.constant([7, 8, 9, 10, 11, 12], shape=[3, 2]) => [[7. 8.]
[9. 10.]
[11. 12.]]
c = tf.matmul(a, b) => [[58 64]
[139 154]]
```

##### Args:

:`a`

`Tensor`

of type`float`

,`double`

,`int32`

or`complex64`

.:`b`

`Tensor`

with same type as`a`

.: If`transpose_a`

`True`

,`a`

is transposed before multiplication.: If`transpose_b`

`True`

,`b`

is transposed before multiplication.: If`a_is_sparse`

`True`

,`a`

is treated as a sparse matrix.: If`b_is_sparse`

`True`

,`b`

is treated as a sparse matrix.: Name for the operation (optional).`name`

##### Returns:

A `Tensor`

of the same type as `a`

.

`tf.batch_matmul(x, y, adj_x=None, adj_y=None, name=None)`

Multiplies slices of two tensors in batches.

Multiplies all slices of `Tensor`

`x`

and `y`

(each slice can be
viewed as an element of a batch), and arranges the individual results
in a single output tensor of the same batch size. Each of the
individual slices can optionally be adjointed (to adjoint a matrix
means to transpose and conjugate it) before multiplication by setting
the `adj_x`

or `adj_y`

flag to `True`

, which are by default `False`

.

The input tensors `x`

and `y`

are 3-D or higher with shape `[..., r_x, c_x]`

and `[..., r_y, c_y]`

.

The output tensor is 3-D or higher with shape `[..., r_o, c_o]`

, where:

```
r_o = c_x if adj_x else r_x
c_o = r_y if adj_y else c_y
```

It is computed as:

```
output[..., :, :] = matrix(x[..., :, :]) * matrix(y[..., :, :])
```

##### Args:

: A`x`

`Tensor`

. Must be one of the following types:`half`

,`float32`

,`float64`

,`int32`

,`complex64`

,`complex128`

. 3-D or higher with shape`[..., r_x, c_x]`

.: A`y`

`Tensor`

. Must have the same type as`x`

. 3-D or higher with shape`[..., r_y, c_y]`

.: An optional`adj_x`

`bool`

. Defaults to`False`

. If`True`

, adjoint the slices of`x`

. Defaults to`False`

.: An optional`adj_y`

`bool`

. Defaults to`False`

. If`True`

, adjoint the slices of`y`

. Defaults to`False`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `x`

.
3-D or higher with shape `[..., r_o, c_o]`

`tf.matrix_determinant(input, name=None)`

Calculates the determinant of a square matrix.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

. A tensor of shape`[M, M]`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

.
A scalar, equal to the determinant of the input.

`tf.batch_matrix_determinant(input, name=None)`

Calculates the determinants for a batch of square matrices.

The input is a tensor of shape `[..., M, M]`

whose inner-most 2 dimensions
form square matrices. The output is a 1-D tensor containing the determinants
for all input submatrices `[..., :, :]`

.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

. Shape is`[..., M, M]`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. Shape is `[...]`

.

`tf.matrix_inverse(input, adjoint=None, name=None)`

Calculates the inverse of a square invertible matrix or its adjoint (conjugate

transpose).

The op uses LU decomposition with partial pivoting to compute the inverse.

If the matrix is not invertible there is no guarantee what the op does. It may detect the condition and raise an exception or it may simply return a garbage result.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[M, M]`

.: An optional`adjoint`

`bool`

. Defaults to`False`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

.
Shape is `[M, M]`

. If `adjoint`

is `False`

then `output`

contains the
matrix inverse of `input`

. If `adjoint`

is `True`

then `output`

contains the
matrix inverse of the adjoint of `input`

.

`tf.batch_matrix_inverse(input, adjoint=None, name=None)`

Calculates the inverse of square invertible matrices or their adjoints

(conjugate transposes).

The input is a tensor of shape `[..., M, M]`

whose inner-most 2 dimensions
form square matrices. The output is a tensor of the same shape as the input
containing the inverse for all input submatrices `[..., :, :]`

.

The op uses LU decomposition with partial pivoting to compute the inverses.

If a matrix is not invertible there is no guarantee what the op does. It may detect the condition and raise an exception or it may simply return a garbage result.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[..., M, M]`

.: An optional`adjoint`

`bool`

. Defaults to`False`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. Shape is `[..., M, M]`

.

`tf.cholesky(input, name=None)`

Calculates the Cholesky decomposition of a square matrix.

The input has to be symmetric and positive definite. Only the lower-triangular part of the input will be used for this operation. The upper-triangular part will not be read.

The result is the lower-triangular matrix of the Cholesky decomposition of the
input, `L`

, so that `input = L L^*`

.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[M, M]`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. Shape is `[M, M]`

.

`tf.batch_cholesky(input, name=None)`

Calculates the Cholesky decomposition of a batch of square matrices.

The input is a tensor of shape `[..., M, M]`

whose inner-most 2 dimensions
form square matrices, with the same constraints as the single matrix Cholesky
decomposition above. The output is a tensor of the same shape as the input
containing the Cholesky decompositions for all input submatrices `[..., :, :]`

.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[..., M, M]`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. Shape is `[..., M, M]`

.

`tf.cholesky_solve(chol, rhs, name=None)`

Solve linear equations `A X = RHS`

, given Cholesky factorization of `A`

.

```
# Solve one system of linear equations (K = 1).
A = [[3, 1], [1, 3]]
RHS = [[2], [22]] # shape 2 x 1
chol = tf.cholesky(A)
X = tf.cholesky_solve(chol, RHS)
# tf.matmul(A, X) ~ RHS
X[:, 0] # Solution to the linear system A x = RHS[:, 0]
# Solve five systems of linear equations (K = 5).
A = [[3, 1], [1, 3]]
RHS = [[1, 2, 3, 4, 5], [11, 22, 33, 44, 55]] # shape 2 x 5
...
X[:, 2] # Solution to the linear system A x = RHS[:, 2]
```

##### Args:

: A`chol`

`Tensor`

. Must be`float32`

or`float64`

, shape is`[M, M]`

. Cholesky factorization of`A`

, e.g.`chol = tf.cholesky(A)`

. For that reason, only the lower triangular part (including the diagonal) of`chol`

is used. The strictly upper part is assumed to be zero and not accessed.: A`rhs`

`Tensor`

, same type as`chol`

, shape is`[M, K]`

, designating`K`

systems of linear equations.: A name to give this`name`

`Op`

. Defaults to`cholesky_solve`

.

##### Returns:

Solution to `A X = RHS`

, shape `[M, K]`

. The solutions to the `K`

systems.

`tf.batch_cholesky_solve(chol, rhs, name=None)`

Solve batches of linear eqns `A X = RHS`

, given Cholesky factorizations.

```
# Solve one linear system (K = 1) for every member of the length 10 batch.
A = ... # shape 10 x 2 x 2
RHS = ... # shape 10 x 2 x 1
chol = tf.batch_cholesky(A) # shape 10 x 2 x 2
X = tf.batch_cholesky_solve(chol, RHS) # shape 10 x 2 x 1
# tf.matmul(A, X) ~ RHS
X[3, :, 0] # Solution to the linear system A[3, :, :] x = RHS[3, :, 0]
# Solve five linear systems (K = 5) for every member of the length 10 batch.
A = ... # shape 10 x 2 x 2
RHS = ... # shape 10 x 2 x 5
...
X[3, :, 2] # Solution to the linear system A[3, :, :] x = RHS[3, :, 2]
```

##### Args:

: A`chol`

`Tensor`

. Must be`float32`

or`float64`

, shape is`[..., M, M]`

. Cholesky factorization of`A`

, e.g.`chol = tf.batch_cholesky(A)`

. For that reason, only the lower triangular parts (including the diagonal) of the last two dimensions of`chol`

are used. The strictly upper part is assumed to be zero and not accessed.: A`rhs`

`Tensor`

, same type as`chol`

, shape is`[..., M, K]`

.: A name to give this`name`

`Op`

. Defaults to`batch_cholesky_solve`

.

##### Returns:

Solution to `A x = rhs`

, shape `[..., M, K]`

.

`tf.self_adjoint_eig(input, name=None)`

Calculates the Eigen Decomposition of a square Self-Adjoint matrix.

Only the lower-triangular part of the input will be used in this case. The upper-triangular part will not be read.

The result is a M+1 x M matrix whose first row is the eigenvalues, and subsequent rows are eigenvectors.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[M, M]`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. Shape is `[M+1, M]`

.

`tf.batch_self_adjoint_eig(input, name=None)`

Calculates the Eigen Decomposition of a batch of square self-adjoint matrices.

The input is a tensor of shape `[..., M, M]`

whose inner-most 2 dimensions
form square matrices, with the same constraints as the single matrix
SelfAdjointEig.

The result is a '[..., M+1, M] matrix with [..., 0,:] containing the eigenvalues, and subsequent [...,1:, :] containing the eigenvectors.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[..., M, M]`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

. Shape is `[..., M+1, M]`

.

`tf.matrix_solve(matrix, rhs, adjoint=None, name=None)`

Solves a system of linear equations. Checks for invertibility.

##### Args:

: A`matrix`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[M, M]`

.: A`rhs`

`Tensor`

. Must have the same type as`matrix`

. Shape is`[M, K]`

.: An optional`adjoint`

`bool`

. Defaults to`False`

. Boolean indicating whether to solve with`matrix`

or its adjoint.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `matrix`

.
Shape is `[M, K]`

. If `adjoint`

is `False`

then `output`

that solves
`matrix`

* output = rhs. If adjoint is True then output that solves
adjoint(matrix) *

`output`

= `rhs`

.`tf.batch_matrix_solve(matrix, rhs, adjoint=None, name=None)`

Solves systems of linear equations. Checks for invertibility.

Matrix is a tensor of shape `[..., M, M]`

whose inner-most 2 dimensions
form square matrices. Rhs is a tensor of shape
`[..., M, K]`

. The output is a tensor shape `[..., M, K]`

. If `adjoint`

is `False`

then each output
matrix satisfies `matrix[..., :, :] * output[..., :, :] = rhs[..., :, :]`

.
If `adjoint`

is `True`

then each output
matrix satisfies `adjoint(matrix[..., :, :]) * output[..., :, :] = rhs[..., :, :]`

.

##### Args:

: A`matrix`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[..., M, M]`

.: A`rhs`

`Tensor`

. Must have the same type as`matrix`

. Shape is`[..., M, K]`

.: An optional`adjoint`

`bool`

. Defaults to`False`

. Boolean indicating whether to solve with`matrix`

or its (block-wise) adjoint.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `matrix`

. Shape is `[..., M, K]`

.

`tf.matrix_triangular_solve(matrix, rhs, lower=None, adjoint=None, name=None)`

Solves a system of linear equations with an upper or lower triangular matrix by

backsubstitution.

`matrix`

is a matrix of shape `[M, M]`

. If `lower`

is `True`

then the strictly
upper triangular part of `matrix`

is assumed to be zero and not accessed.
If `lower`

is False then the strictly lower triangular part of `matrix`

is
assumed to be zero and not accessed.
`rhs`

is a matrix of shape [M, K]`.

The output is a matrix of shape `[M, K]`

. If `adjoint`

is `False`

the output
satisfies the matrix equation `matrix`

* output = rhs.
If adjoint is False then output satisfies the matrix equation
matrix *

`output`

= `rhs`

.
If `adjoint`

is `True`

then `output`

satisfies the matrix equation
`adjoint(matrix)`

* `output`

= `rhs`

.##### Args:

: A`matrix`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[M, M]`

.: A`rhs`

`Tensor`

. Must have the same type as`matrix`

. Shape is`[M, K]`

.: An optional`lower`

`bool`

. Defaults to`True`

. Boolean indicating whether`matrix`

is lower or upper triangular: An optional`adjoint`

`bool`

. Defaults to`False`

. Boolean indicating whether to solve with`matrix`

or its adjoint.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `matrix`

. Shape is `[M, K]`

.

`tf.batch_matrix_triangular_solve(matrix, rhs, lower=None, adjoint=None, name=None)`

Solves systems of linear equations with upper or lower triangular matrices by

backsubstitution.

`matrix`

is a tensor of shape `[..., M, M]`

whose inner-most 2 dimensions form
square matrices. If `lower`

is `True`

then the strictly upper triangular part
of each inner-most matrix is assumed to be zero and not accessed.
If `lower`

is False then the strictly lower triangular part of each inner-most
matrix is assumed to be zero and not accessed.
`rhs`

is a tensor of shape [..., M, K]`.

The output is a tensor of shape `[..., M, K]`

. If `adjoint`

is `True`

then the
innermost matrices in output`satisfy matrix equations`

matrix[..., :, :] * output[..., :, :] = rhs[..., :, :] .
IfadjointisFalsethen the strictly then the innermost matrices inoutputsatisfy matrix equationsadjoint(matrix[..., i, k]) * output[..., k, j] = rhs[..., i, j]`.

##### Args:

: A`matrix`

`Tensor`

. Must be one of the following types:`float64`

,`float32`

. Shape is`[..., M, M]`

.: A`rhs`

`Tensor`

. Must have the same type as`matrix`

. Shape is`[..., M, K]`

.: An optional`lower`

`bool`

. Defaults to`True`

. Boolean indicating whether the innermost matrices in`matrix`

are lower or upper triangular.: An optional`adjoint`

`bool`

. Defaults to`False`

. Boolean indicating whether to solve with`matrix`

or its (block-wise) adjoint.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `matrix`

. Shape is `[..., M, K]`

.

`tf.matrix_solve_ls(matrix, rhs, l2_regularizer=0.0, fast=True, name=None)`

Solves a linear least-squares problem.

Below we will use the following notation
`matrix`

=\(A \in \Re^{m \times n}\),
`rhs`

=\(B \in \Re^{m \times k}\),
`output`

=\(X \in \Re^{n \times k}\),
`l2_regularizer`

=\(\lambda\).

If `fast`

is `True`

, then the solution is computed by solving the normal
equations using Cholesky decomposition. Specifically, if \(m \ge n\) then
\(X = (A^T A + \lambda I)^{-1} A^T B\), which solves the regularized
least-squares problem \(X = \mathrm{argmin}*{Z \in \Re^{n \times k}}
||A Z - B||_F^2 + \lambda ||Z||_F^2\). If \(m \lt n\) then output is
computed as \(X = A^T (A A^T + \lambda I)^{-1} B\),
which (for \(\lambda = 0\)) is the minimum-norm solution to the
under-determined linear system, i.e.
\(X = \mathrm{argmin}*{Z \in \Re^{n \times k}} ||Z||

*F^2 \), subject to \(A Z = B\). Notice that the fast path is only numerically stable when \(A\) is numerically full rank and has a condition number \(\mathrm{cond}(A) \lt \frac{1}{\sqrt{\epsilon*{mach}}}\) or \(\lambda\) is sufficiently large.

If `fast`

is `False`

then the solution is computed using the rank revealing
QR decomposition with column pivoting. This will always compute a
least-squares solution that minimizes the residual norm
\(||A X - B||_F^2 \), even when \(A\) is rank deficient or
ill-conditioned. Notice: The current version does not compute a minimum norm
solution. If `fast`

is `False`

then `l2_regularizer`

is ignored.

##### Args:

: 2-D`matrix`

`Tensor`

of shape`[M, N]`

.: 2-D`rhs`

`Tensor`

of shape is`[M, K]`

.: 0-D`l2_regularizer`

`double`

`Tensor`

. Ignored if`fast=False`

.: bool. Defaults to`fast`

`True`

.: string, optional name of the operation.`name`

##### Returns:

: Matrix of shape`output`

`[N, K]`

containing the matrix that solves`matrix * output = rhs`

in the least-squares sense.

`tf.batch_matrix_solve_ls(matrix, rhs, l2_regularizer=0.0, fast=True, name=None)`

Solves multiple linear least-squares problems.

`matrix`

is a tensor of shape `[..., M, N]`

whose inner-most 2 dimensions
form `M`

-by-`N`

matrices. Rhs is a tensor of shape `[..., M, K]`

whose
inner-most 2 dimensions form `M`

-by-`K`

matrices. The computed output is a
`Tensor`

of shape `[..., N, K]`

whose inner-most 2 dimensions form `M`

-by-`K`

matrices that solve the equations
`matrix[..., :, :] * output[..., :, :] = rhs[..., :, :]`

in the least squares
sense.

Below we will use the following notation for each pair of matrix and right-hand sides in the batch:

`matrix`

=\(A \in \Re^{m \times n}\),
`rhs`

=\(B \in \Re^{m \times k}\),
`output`

=\(X \in \Re^{n \times k}\),
`l2_regularizer`

=\(\lambda\).

If `fast`

is `True`

, then the solution is computed by solving the normal
equations using Cholesky decomposition. Specifically, if \(m \ge n\) then
\(X = (A^T A + \lambda I)^{-1} A^T B\), which solves the least-squares
problem \(X = \mathrm{argmin}*{Z \in \Re^{n \times k}} ||A Z - B||_F^2 +
\lambda ||Z||_F^2\). If \(m \lt n\) then output is computed as
\(X = A^T (A A^T + \lambda I)^{-1} B\), which (for \(\lambda = 0\)) is
the minimum-norm solution to the under-determined linear system, i.e.
\(X = \mathrm{argmin}*{Z \in \Re^{n \times k}} ||Z||

*F^2 \), subject to \(A Z = B\). Notice that the fast path is only numerically stable when \(A\) is numerically full rank and has a condition number \(\mathrm{cond}(A) \lt \frac{1}{\sqrt{\epsilon*{mach}}}\) or\(\lambda\) is sufficiently large.

If `fast`

is `False`

an algorithm based on the numerically robust complete
orthogonal decomposition is used. This computes the minimum-norm
least-squares solution, even when \(A\) is rank deficient. This path is
typically 6-7 times slower than the fast path. If `fast`

is `False`

then
`l2_regularizer`

is ignored.

##### Args:

:`matrix`

`Tensor`

of shape`[..., M, N]`

.:`rhs`

`Tensor`

of shape`[..., M, K]`

.: 0-D`l2_regularizer`

`double`

`Tensor`

. Ignored if`fast=False`

.: bool. Defaults to`fast`

`True`

.: string, optional name of the operation.`name`

##### Returns:

:`output`

`Tensor`

of shape`[..., N, K]`

whose inner-most 2 dimensions form`M`

-by-`K`

matrices that solve the equations`matrix[..., :, :] * output[..., :, :] = rhs[..., :, :]`

in the least squares sense.

## Complex Number Functions

TensorFlow provides several operations that you can use to add complex number functions to your graph.

`tf.complex(real, imag, name=None)`

Converts two real numbers to a complex number.

Given a tensor `real`

representing the real part of a complex number, and a
tensor `imag`

representing the imaginary part of a complex number, this
operation returns complex numbers elementwise of the form (a + bj), where
*a* represents the `real`

part and *b* represents the `imag`

part.

The input tensors `real`

and `imag`

must have the same shape.

For example:

```
# tensor 'real' is [2.25, 3.25]
# tensor `imag` is [4.75, 5.75]
tf.complex(real, imag) ==> [[2.25 + 4.75j], [3.25 + 5.75j]]
```

##### Args:

: A`real`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`imag`

`Tensor`

. Must have the same type as`real`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

or `complex128`

.

`tf.complex_abs(x, name=None)`

Computes the complex absolute value of a tensor.

Given a tensor `x`

of complex numbers, this operation returns a tensor of type
`float`

or `double`

that is the absolute value of each element in `x`

. All
elements in `x`

must be complex numbers of the form \(a + bj\). The
absolute value is computed as \( \sqrt{a^2 + b^2}\).

For example:

```
# tensor 'x' is [[-2.25 + 4.75j], [-3.25 + 5.75j]]
tf.complex_abs(x) ==> [5.25594902, 6.60492229]
```

##### Args:

: A`x`

`Tensor`

of type`complex64`

or`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `float32`

or `float64`

.

`tf.conj(input, name=None)`

Returns the complex conjugate of a complex number.

Given a tensor `input`

of complex numbers, this operation returns a tensor of
complex numbers that are the complex conjugate of each element in `input`

. The
complex numbers in `input`

must be of the form \(a + bj\), where *a* is the
real part and *b* is the imaginary part.

The complex conjugate returned by this operation is of the form \(a - bj\).

For example:

```
# tensor 'input' is [-2.25 + 4.75j, 3.25 + 5.75j]
tf.conj(input) ==> [-2.25 - 4.75j, 3.25 - 5.75j]
```

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `input`

.

`tf.imag(input, name=None)`

Returns the imaginary part of a complex number.

Given a tensor `input`

of complex numbers, this operation returns a tensor of
type `float`

or `double`

that is the imaginary part of each element in
`input`

. All elements in `input`

must be complex numbers of the form (a +
bj), where *a* is the real part and *b* is the imaginary part returned by
this operation.

For example:

```
# tensor 'input' is [-2.25 + 4.75j, 3.25 + 5.75j]
tf.imag(input) ==> [4.75, 5.75]
```

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`complex64`

,`complex128`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `float`

or `double`

.

`tf.real(input, name=None)`

Returns the real part of a complex number.

Given a tensor `input`

of complex numbers, this operation returns a tensor of
type `float`

or `double`

that is the real part of each element in `input`

.
All elements in `input`

must be complex numbers of the form (a + bj),
where *a* is the real part returned by this operation and *b* is the
imaginary part.

For example:

```
# tensor 'input' is [-2.25 + 4.75j, 3.25 + 5.75j]
tf.real(input) ==> [-2.25, 3.25]
```

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`complex64`

,``complex128`.`

: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `float`

or `double`

.

`tf.fft(input, name=None)`

Compute the 1-dimensional discrete Fourier Transform.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 vector.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

. The 1D Fourier Transform of `input`

.

`tf.ifft(input, name=None)`

Compute the inverse 1-dimensional discrete Fourier Transform.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 vector.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
The inverse 1D Fourier Transform of `input`

.

`tf.fft2d(input, name=None)`

Compute the 2-dimensional discrete Fourier Transform.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 matrix.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

. The 2D Fourier Transform of `input`

.

`tf.ifft2d(input, name=None)`

Compute the inverse 2-dimensional discrete Fourier Transform.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 matrix.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
The inverse 2D Fourier Transform of `input`

.

`tf.fft3d(input, name=None)`

Compute the 3-dimensional discrete Fourier Transform.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 3-D tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

. The 3D Fourier Transform of `input`

.

`tf.ifft3d(input, name=None)`

Compute the inverse 3-dimensional discrete Fourier Transform.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 3-D tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
The inverse 3D Fourier Transform of `input`

.

`tf.batch_fft(input, name=None)`

Compute the 1-dimensional discrete Fourier Transform over the inner-most

dimension of `input`

.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
A complex64 tensor of the same shape as `input`

. The inner-most
dimension of `input`

is replaced with its 1D Fourier Transform.

`tf.batch_ifft(input, name=None)`

Compute the inverse 1-dimensional discrete Fourier Transform over the inner-most

dimension of `input`

.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
A complex64 tensor of the same shape as `input`

. The inner-most
dimension of `input`

is replaced with its inverse 1D Fourier Transform.

`tf.batch_fft2d(input, name=None)`

Compute the 2-dimensional discrete Fourier Transform over the inner-most

2 dimensions of `input`

.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
A complex64 tensor of the same shape as `input`

. The inner-most 2
dimensions of `input`

are replaced with their 2D Fourier Transform.

`tf.batch_ifft2d(input, name=None)`

Compute the inverse 2-dimensional discrete Fourier Transform over the inner-most

2 dimensions of `input`

.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
A complex64 tensor of the same shape as `input`

. The inner-most 2
dimensions of `input`

are replaced with their inverse 2D Fourier Transform.

`tf.batch_fft3d(input, name=None)`

Compute the 3-dimensional discrete Fourier Transform over the inner-most 3

dimensions of `input`

.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
A complex64 tensor of the same shape as `input`

. The inner-most 3
dimensions of `input`

are replaced with their 3D Fourier Transform.

`tf.batch_ifft3d(input, name=None)`

Compute the inverse 3-dimensional discrete Fourier Transform over the inner-most

3 dimensions of `input`

.

##### Args:

: A`input`

`Tensor`

of type`complex64`

. A complex64 tensor.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `complex64`

.
A complex64 tensor of the same shape as `input`

. The inner-most 3
dimensions of `input`

are replaced with their inverse 3D Fourier Transform.

## Reduction

TensorFlow provides several operations that you can use to perform common math computations that reduce various dimensions of a tensor.

`tf.reduce_sum(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the sum of elements across dimensions of a tensor.

Reduces `input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

If `reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

For example:

```
# 'x' is [[1, 1, 1]
# [1, 1, 1]]
tf.reduce_sum(x) ==> 6
tf.reduce_sum(x, 0) ==> [2, 2, 2]
tf.reduce_sum(x, 1) ==> [3, 3]
tf.reduce_sum(x, 1, keep_dims=True) ==> [[3], [3]]
tf.reduce_sum(x, [0, 1]) ==> 6
```

##### Args:

: The tensor to reduce. Should have numeric type.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.reduce_prod(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the product of elements across dimensions of a tensor.

Reduces `input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

If `reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

##### Args:

: The tensor to reduce. Should have numeric type.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.reduce_min(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the minimum of elements across dimensions of a tensor.

Reduces `input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

If `reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

##### Args:

: The tensor to reduce. Should have numeric type.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.reduce_max(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the maximum of elements across dimensions of a tensor.

`input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

`reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

##### Args:

: The tensor to reduce. Should have numeric type.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.reduce_mean(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the mean of elements across dimensions of a tensor.

`input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

`reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

For example:

```
# 'x' is [[1., 1.]
# [2., 2.]]
tf.reduce_mean(x) ==> 1.5
tf.reduce_mean(x, 0) ==> [1.5, 1.5]
tf.reduce_mean(x, 1) ==> [1., 2.]
```

##### Args:

: The tensor to reduce. Should have numeric type.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.reduce_all(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the "logical and" of elements across dimensions of a tensor.

`input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

`reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

For example:

```
# 'x' is [[True, True]
# [False, False]]
tf.reduce_all(x) ==> False
tf.reduce_all(x, 0) ==> [False, False]
tf.reduce_all(x, 1) ==> [True, False]
```

##### Args:

: The boolean tensor to reduce.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.reduce_any(input_tensor, reduction_indices=None, keep_dims=False, name=None)`

Computes the "logical or" of elements across dimensions of a tensor.

`input_tensor`

along the dimensions given in `reduction_indices`

.
Unless `keep_dims`

is true, the rank of the tensor is reduced by 1 for each
entry in `reduction_indices`

. If `keep_dims`

is true, the reduced dimensions
are retained with length 1.

`reduction_indices`

has no entries, all dimensions are reduced, and a
tensor with a single element is returned.

For example:

```
# 'x' is [[True, True]
# [False, False]]
tf.reduce_any(x) ==> True
tf.reduce_any(x, 0) ==> [True, True]
tf.reduce_any(x, 1) ==> [True, False]
```

##### Args:

: The boolean tensor to reduce.`input_tensor`

: The dimensions to reduce. If`reduction_indices`

`None`

(the default), reduces all dimensions.: If true, retains reduced dimensions with length 1.`keep_dims`

: A name for the operation (optional).`name`

##### Returns:

The reduced tensor.

`tf.accumulate_n(inputs, shape=None, tensor_dtype=None, name=None)`

Returns the element-wise sum of a list of tensors.

Optionally, pass `shape`

and `tensor_dtype`

for shape and type checking,
otherwise, these are inferred.

For example:

```
# tensor 'a' is [[1, 2], [3, 4]]
# tensor `b` is [[5, 0], [0, 6]]
tf.accumulate_n([a, b, a]) ==> [[7, 4], [6, 14]]
# Explicitly pass shape and type
tf.accumulate_n([a, b, a], shape=[2, 2], tensor_dtype=tf.int32)
==> [[7, 4], [6, 14]]
```

##### Args:

: A list of`inputs`

`Tensor`

objects, each with same shape and type.: Shape of elements of`shape`

`inputs`

.: The type of`tensor_dtype`

`inputs`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of same shape and type as the elements of `inputs`

.

##### Raises:

: If`ValueError`

`inputs`

don't all have same shape and dtype or the shape cannot be inferred.

## Segmentation

TensorFlow provides several operations that you can use to perform common
math computations on tensor segments.
Here a segmentation is a partitioning of a tensor along
the first dimension, i.e. it defines a mapping from the first dimension onto
`segment_ids`

. The `segment_ids`

tensor should be the size of
the first dimension, `d0`

, with consecutive IDs in the range `0`

to `k`

,
where `k<d0`

.
In particular, a segmentation of a matrix tensor is a mapping of rows to
segments.

For example:

```
c = tf.constant([[1,2,3,4], [-1,-2,-3,-4], [5,6,7,8]])
tf.segment_sum(c, tf.constant([0, 0, 1]))
==> [[0 0 0 0]
[5 6 7 8]]
```

`tf.segment_sum(data, segment_ids, name=None)`

Computes the sum along segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Computes a tensor such that
\(output_i = \sum_j data_j\) where sum is over `j`

such
that `segment_ids[j] == i`

.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int64`

,`int32`

,`uint8`

,`uint16`

,`int16`

,`int8`

,`complex64`

,`complex128`

,`qint8`

,`quint8`

,`qint32`

,`half`

.: A`segment_ids`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

. A 1-D tensor whose rank is equal to the rank of`data`

's first dimension. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.segment_prod(data, segment_ids, name=None)`

Computes the product along segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Computes a tensor such that
\(output_i = \prod_j data_j\) where the product is over `j`

such
that `segment_ids[j] == i`

.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int64`

,`int32`

,`uint8`

,`uint16`

,`int16`

,`int8`

,`complex64`

,`complex128`

,`qint8`

,`quint8`

,`qint32`

,`half`

.: A`segment_ids`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

. A 1-D tensor whose rank is equal to the rank of`data`

's first dimension. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.segment_min(data, segment_ids, name=None)`

Computes the minimum along segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Computes a tensor such that
\(output_i = \min_j(data_j)\) where `min`

is over `j`

such
that `segment_ids[j] == i`

.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`uint8`

,`int16`

,`int8`

,`uint16`

,`half`

.: A`segment_ids`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

. A 1-D tensor whose rank is equal to the rank of`data`

's first dimension. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.segment_max(data, segment_ids, name=None)`

Computes the maximum along segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Computes a tensor such that
\(output_i = \max_j(data_j)\) where `max`

is over `j`

such
that `segment_ids[j] == i`

.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`uint8`

,`int16`

,`int8`

,`uint16`

,`half`

.: A`segment_ids`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

. A 1-D tensor whose rank is equal to the rank of`data`

's first dimension. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

`Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.segment_mean(data, segment_ids, name=None)`

Computes the mean along segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Computes a tensor such that
\(output_i = \frac{\sum_j data_j}{N}\) where `mean`

is
over `j`

such that `segment_ids[j] == i`

and `N`

is the total number of
values summed.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`uint8`

,`int16`

,`int8`

,`uint16`

,`half`

.: A`segment_ids`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

. A 1-D tensor whose rank is equal to the rank of`data`

's first dimension. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

`Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.unsorted_segment_sum(data, segment_ids, num_segments, name=None)`

Computes the sum along segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Computes a tensor such that
\(output_i = \sum_j data_j\) where sum is over `j`

such
that `segment_ids[j] == i`

. Unlike `SegmentSum`

, `segment_ids`

need not be sorted and need not cover all values in the full
range of valid values.

If the sum is empty for a given segment ID `i`

, `output[i] = 0`

.

`num_segments`

should equal the number of distinct segment IDs.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int64`

,`int32`

,`uint8`

,`uint16`

,`int16`

,`int8`

,`complex64`

,`complex128`

,`qint8`

,`quint8`

,`qint32`

,`half`

.: A`segment_ids`

`Tensor`

. Must be one of the following types:`int32`

,`int64`

. A 1-D tensor whose rank is equal to the rank of`data`

's first dimension.: A`num_segments`

`Tensor`

of type`int32`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `num_segments`

.

`tf.sparse_segment_sum(data, indices, segment_ids, name=None)`

Computes the sum along sparse segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Like `SegmentSum`

, but `segment_ids`

can have rank less than `data`

's first
dimension, selecting a subset of dimension 0, specified by `indices`

.

For example:

```
c = tf.constant([[1,2,3,4], [-1,-2,-3,-4], [5,6,7,8]])
# Select two rows, one segment.
tf.sparse_segment_sum(c, tf.constant([0, 1]), tf.constant([0, 0]))
==> [[0 0 0 0]]
# Select two rows, two segment.
tf.sparse_segment_sum(c, tf.constant([0, 1]), tf.constant([0, 1]))
==> [[ 1 2 3 4]
[-1 -2 -3 -4]]
# Select all rows, two segments.
tf.sparse_segment_sum(c, tf.constant([0, 1, 2]), tf.constant([0, 0, 1]))
==> [[0 0 0 0]
[5 6 7 8]]
# Which is equivalent to:
tf.segment_sum(c, tf.constant([0, 0, 1]))
```

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int32`

,`int64`

,`uint8`

,`int16`

,`int8`

,`uint16`

,`half`

.: A`indices`

`Tensor`

of type`int32`

. A 1-D tensor. Has same rank as`segment_ids`

.: A`segment_ids`

`Tensor`

of type`int32`

. A 1-D tensor. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

`Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.sparse_segment_mean(data, indices, segment_ids, name=None)`

Computes the mean along sparse segments of a tensor.

Read the section on Segmentation for an explanation of segments.

Like `SegmentMean`

, but `segment_ids`

can have rank less than `data`

's first
dimension, selecting a subset of dimension 0, specified by `indices`

.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`indices`

`Tensor`

of type`int32`

. A 1-D tensor. Has same rank as`segment_ids`

.: A`segment_ids`

`Tensor`

of type`int32`

. A 1-D tensor. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

`Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

`tf.sparse_segment_sqrt_n(data, indices, segment_ids, name=None)`

Computes the sum along sparse segments of a tensor divided by the sqrt of N.

N is the size of the segment being reduced.

Read the section on Segmentation for an explanation of segments.

##### Args:

: A`data`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

.: A`indices`

`Tensor`

of type`int32`

. A 1-D tensor. Has same rank as`segment_ids`

.: A`segment_ids`

`Tensor`

of type`int32`

. A 1-D tensor. Values should be sorted and can be repeated.: A name for the operation (optional).`name`

##### Returns:

`Tensor`

. Has the same type as `data`

.
Has same shape as data, except for dimension 0 which
has size `k`

, the number of segments.

## Sequence Comparison and Indexing

TensorFlow provides several operations that you can use to add sequence comparison and index extraction to your graph. You can use these operations to determine sequence differences and determine the indexes of specific values in a tensor.

`tf.argmin(input, dimension, name=None)`

Returns the index with the smallest value across dimensions of a tensor.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int64`

,`int32`

,`uint8`

,`uint16`

,`int16`

,`int8`

,`complex64`

,`complex128`

,`qint8`

,`quint8`

,`qint32`

,`half`

.: A`dimension`

`Tensor`

of type`int32`

. int32, 0 <= dimension < rank(input). Describes which dimension of the input Tensor to reduce across. For vectors, use dimension = 0.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `int64`

.

`tf.argmax(input, dimension, name=None)`

Returns the index with the largest value across dimensions of a tensor.

##### Args:

: A`input`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

,`int64`

,`int32`

,`uint8`

,`uint16`

,`int16`

,`int8`

,`complex64`

,`complex128`

,`qint8`

,`quint8`

,`qint32`

,`half`

.: A`dimension`

`Tensor`

of type`int32`

. int32, 0 <= dimension < rank(input). Describes which dimension of the input Tensor to reduce across. For vectors, use dimension = 0.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `int64`

.

`tf.listdiff(x, y, name=None)`

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:

: A`x`

`Tensor`

. 1-D. Values to keep.: A`y`

`Tensor`

. Must have the same type as`x`

. 1-D. Values to remove.: A name for the operation (optional).`name`

##### Returns:

A tuple of `Tensor`

objects (out, idx).

: A`out`

`Tensor`

. Has the same type as`x`

. 1-D. Values present in`x`

but not in`y`

.: A`idx`

`Tensor`

of type`int32`

. 1-D. Positions of`x`

values preserved in`out`

.

`tf.where(input, name=None)`

Returns locations of true values in a boolean tensor.

This operation returns the coordinates of true elements in `input`

. 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
`input`

. 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]]
# `input` 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]]
```

##### Args:

: A`input`

`Tensor`

of type`bool`

.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `int64`

.

`tf.unique(x, name=None)`

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`

. 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 = unique(x)
y ==> [1, 2, 4, 7, 8]
idx ==> [0, 0, 1, 2, 2, 2, 3, 4, 4]
```

##### Args:

: A`x`

`Tensor`

. 1-D.: A name for the operation (optional).`name`

##### Returns:

A tuple of `Tensor`

objects (y, idx).

: A`y`

`Tensor`

. Has the same type as`x`

. 1-D.: A`idx`

`Tensor`

of type`int32`

. 1-D.

`tf.edit_distance(hypothesis, truth, normalize=True, name='edit_distance')`

Computes the Levenshtein distance between sequences.

This operation takes variable-length sequences (`hypothesis`

and `truth`

),
each provided as a `SparseTensor`

, and computes the Levenshtein distance.
You can normalize the edit distance by length of `truth`

by setting
`normalize`

to true.

For example, given the following input:

```
# 'hypothesis' is a tensor of shape `[2, 1]` with variable-length values:
# (0,0) = ["a"]
# (1,0) = ["b"]
hypothesis = tf.SparseTensor(
[[0, 0, 0],
[1, 0, 0]],
["a", "b"]
(2, 1, 1))
# 'truth' is a tensor of shape `[2, 2]` with variable-length values:
# (0,0) = []
# (0,1) = ["a"]
# (1,0) = ["b", "c"]
# (1,1) = ["a"]
truth = tf.SparseTensor(
[[0, 1, 0],
[1, 0, 0],
[1, 0, 1],
[1, 1, 0]]
["a", "b", "c", "a"],
(2, 2, 2))
normalize = True
```

This operation would return the following:

```
# 'output' is a tensor of shape `[2, 2]` with edit distances normalized
# by 'truth' lengths.
output ==> [[inf, 1.0], # (0,0): no truth, (0,1): no hypothesis
[0.5, 1.0]] # (1,0): addition, (1,1): no hypothesis
```

##### Args:

: A`hypothesis`

`SparseTensor`

containing hypothesis sequences.: A`truth`

`SparseTensor`

containing truth sequences.: A`normalize`

`bool`

. If`True`

, normalizes the Levenshtein distance by length of`truth.`

: A name for the operation (optional).`name`

##### Returns:

A dense `Tensor`

with rank `R - 1`

, where R is the rank of the
`SparseTensor`

inputs `hypothesis`

and `truth`

.

##### Raises:

: If either`TypeError`

`hypothesis`

or`truth`

are not a`SparseTensor`

.

`tf.invert_permutation(x, name=None)`

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:

: A`x`

`Tensor`

of type`int32`

. 1-D.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

of type `int32`

. 1-D.

## Other Functions and Classes

`tf.scalar_mul(scalar, x)`

Multiplies a scalar times a `Tensor`

or `IndexedSlices`

object.

Intended for use in gradient code which might deal with `IndexedSlices`

objects, which are easy to multiply by a scalar but more expensive to
multiply with arbitrary tensors.

##### Args:

: A 0-D scalar`scalar`

`Tensor`

. Must have known shape.: A`x`

`Tensor`

or`IndexedSlices`

to be scaled.

##### Returns:

`scalar * x`

of the same type (`Tensor`

or `IndexedSlices`

) as `x`

.

##### Raises:

: if scalar is not a 0-D`ValueError`

`scalar`

.

`tf.sparse_segment_sqrt_n_grad(grad, indices, segment_ids, output_dim0, name=None)`

Computes gradients for SparseSegmentSqrtN.

Returns tensor "output" with same shape as grad, except for dimension 0 whose value is output_dim0.

##### Args:

: A`grad`

`Tensor`

. Must be one of the following types:`float32`

,`float64`

. gradient propagated to the SparseSegmentSqrtN op.: A`indices`

`Tensor`

of type`int32`

. indices passed to the corresponding SparseSegmentSqrtN op.: A`segment_ids`

`Tensor`

of type`int32`

. segment_ids passed to the corresponding SparseSegmentSqrtN op.: A`output_dim0`

`Tensor`

of type`int32`

. dimension 0 of "data" passed to SparseSegmentSqrtN op.: A name for the operation (optional).`name`

##### Returns:

A `Tensor`

. Has the same type as `grad`

.