path: root/tensorflow/g3doc/api_docs/python/math_ops.md

<!-- This file is machine generated: DO NOT EDIT! -->

# Math

Note: Functions taking `Tensor` arguments can also take anything accepted by
[`tf.convert_to_tensor`](framework.md#convert_to_tensor).

[TOC]

Note: Elementwise binary operations in TensorFlow follow [numpy-style
broadcasting](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html).

## Arithmetic Operators

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

- - -

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

Returns x + y element-wise.

*NOTE*: `Add` supports broadcasting. `AddN` does not. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `uint8`, `int8`, `int16`, `int32`, `int64`, `complex64`, `complex128`, `string`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.subtract(x, y, name=None)` {#subtract}

Returns x - y element-wise.

*NOTE*: `tf.subtract` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.multiply(x, y, name=None)` {#multiply}

Returns x * y element-wise.

*NOTE*: ``tf.multiply`` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `uint8`, `int8`, `uint16`, `int16`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.scalar_mul(scalar, x)` {#scalar_mul}

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:


*  <b>`scalar`</b>: A 0-D scalar `Tensor`. Must have known shape.
*  <b>`x`</b>: A `Tensor` or `IndexedSlices` to be scaled.

##### Returns:

  `scalar * x` of the same type (`Tensor` or `IndexedSlices`) as `x`.

##### Raises:


*  <b>`ValueError`</b>: if scalar is not a 0-D `scalar`.


- - -

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

Divides x / y elementwise (using Python 2 division operator semantics).

NOTE: Prefer using the Tensor division operator or tf.divide which obey Python
division operator semantics.

This function divides `x` and `y`, forcing Python 2.7 semantics. That is,
if one of `x` or `y` is a float, then the result will be a float.
Otherwise, the output will be an integer type. Flooring semantics are used
for integer division.

##### Args:


*  <b>`x`</b>: `Tensor` numerator of real numeric type.
*  <b>`y`</b>: `Tensor` denominator of real numeric type.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  `x / y` returns the quotient of x and y.


- - -

### `tf.divide(x, y, name=None)` {#divide}

Computes Python style division of `x` by `y`.


- - -

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

Divides x / y elementwise (using Python 3 division operator semantics).

NOTE: Prefer using the Tensor operator or tf.divide which obey Python
division operator semantics.

This function forces Python 3 division operator semantics where all integer
arguments are cast to floating types first.   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:


*  <b>`x`</b>: `Tensor` numerator of numeric type.
*  <b>`y`</b>: `Tensor` denominator of numeric type.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  `x / y` evaluated in floating point.

##### Raises:


*  <b>`TypeError`</b>: If `x` and `y` have different dtypes.


- - -

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

Divides `x / y` elementwise, rounding toward the most negative integer.

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:


*  <b>`x`</b>: `Tensor` numerator of real numeric type.
*  <b>`y`</b>: `Tensor` denominator of real numeric type.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  `x / y` rounded down (except possibly towards zero for negative integers).

##### Raises:


*  <b>`TypeError`</b>: If the inputs are complex.


- - -

### `tf.realdiv(x, y, name=None)` {#realdiv}

Returns x / y element-wise for real types.

If `x` and `y` are reals, this will return the floating-point division.

*NOTE*: `Div` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `uint8`, `int8`, `uint16`, `int16`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.truncatediv(x, y, name=None)` {#truncatediv}

Returns x / y element-wise for integer types.

Truncation designates that negative numbers will round fractional quantities
toward zero. I.e. -7 / 5 = 1. This matches C semantics but it is different
than Python semantics. See `FloorDiv` for a division function that matches
Python Semantics.

*NOTE*: `TruncateDiv` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `uint8`, `int8`, `uint16`, `int16`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.floor_div(x, y, name=None)` {#floor_div}

Returns x // y element-wise.

*NOTE*: `FloorDiv` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `uint8`, `int8`, `uint16`, `int16`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.truncatemod(x, y, name=None)` {#truncatemod}

Returns element-wise remainder of division. This emulates C semantics where

true, this follows C semantics in that the result here is consistent
with a flooring divide. E.g. `floor(x / y) * y + mod(x, y) = x`.

*NOTE*: `Mod` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`, `float32`, `float64`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.floormod(x, y, name=None)` {#floormod}

Returns element-wise remainder of division. When `x < 0` xor `y < 0` is

true, this follows Python semantics in that the result here is consistent
with a flooring divide. E.g. `floor(x / y) * y + mod(x, y) = x`.

*NOTE*: `FloorMod` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`, `float32`, `float64`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

Returns element-wise remainder of division. When `x < 0` xor `y < 0` is

true, this follows Python semantics in that the result here is consistent
with a flooring divide. E.g. `floor(x / y) * y + mod(x, y) = x`.

*NOTE*: `FloorMod` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`, `float32`, `float64`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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:


*  <b>`a`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`, `half`.
    A tensor containing 3-element vectors.
*  <b>`b`</b>: A `Tensor`. Must have the same type as `a`.
    Another tensor, of same type and shape as `a`.
*  <b>`name`</b>: A name for the operation (optional).

##### 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_n}

Adds all input tensors element-wise.

##### Args:


*  <b>`inputs`</b>: A list of `Tensor` objects, each with same shape and type.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of same shape and type as the elements of `inputs`.

##### Raises:


*  <b>`ValueError`</b>: If `inputs` don't all have same shape and dtype or the shape
  cannot be inferred.


- - -

### `tf.abs(x, name=None)` {#abs}

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|\\).

##### Args:


*  <b>`x`</b>: A `Tensor` or `SparseTensor` of type `float32`, `float64`, `int32`, or
    `int64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` or `SparseTensor` the same size and type as `x` with absolute
    values.


- - -

### `tf.negative(x, name=None)` {#negative}

Computes numerical negative value element-wise.

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

##### Args:


*  <b>`x`</b>: A `Tensor` or `SparseTensor`. Must be one of the following types: `half`,
    `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` or `SparseTensor`, respectively. Has the same type as `x`.


- - -

### `tf.sign(x, name=None)` {#sign}

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:


*  <b>`x`</b>: A `Tensor` or `SparseTensor`. Must be one of the following types: `half`,
    `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` or `SparseTensor`, respectively. Has the same type as `x`.


- - -

### `tf.reciprocal(x, name=None)` {#reciprocal}

Computes the reciprocal of x element-wise.

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

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.square(x, name=None)` {#square}

Computes square of x element-wise.

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

##### Args:


*  <b>`x`</b>: A `Tensor` or `SparseTensor`. Must be one of the following types: `half`,
    `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` or `SparseTensor`. Has the same type as `x`.


- - -

### `tf.round(x, name=None)` {#round}

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

Rounds half to even.  Also known as bankers rounding. If you want to round
according to the current system rounding mode use tf::cint.
For example:

```python
# 'a' is [0.9, 2.5, 2.3, 1.5, -4.5]
tf.round(a) ==> [ 1.0, 2.0, 2.0, 2.0, -4.0 ]
```

##### Args:


*  <b>`x`</b>: A `Tensor` of type `float32` or `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of same shape and type as `x`.


- - -

### `tf.sqrt(x, name=None)` {#sqrt}

Computes square root of x element-wise.

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

##### Args:


*  <b>`x`</b>: A `Tensor` or `SparseTensor`. Must be one of the following types: `half`,
    `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` or `SparseTensor`, respectively. Has the same type as `x`.


- - -

### `tf.rsqrt(x, name=None)` {#rsqrt}

Computes reciprocal of square root of x element-wise.

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

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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:


*  <b>`x`</b>: A `Tensor` of type `float32`, `float64`, `int32`, `int64`, `complex64`,
   or `complex128`.
*  <b>`y`</b>: A `Tensor` of type `float32`, `float64`, `int32`, `int64`, `complex64`,
   or `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`.


- - -

### `tf.exp(x, name=None)` {#exp}

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

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.expm1(x, name=None)` {#expm1}

Computes exponential of x - 1 element-wise.

I.e., \\(y = (\exp x) - 1\\).

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.log(x, name=None)` {#log}

Computes natural logarithm of x element-wise.

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

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.log1p(x, name=None)` {#log1p}

Computes natural logarithm of (1 + x) element-wise.

I.e., \\(y = \log_e (1 + x)\\).

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.ceil(x, name=None)` {#ceil}

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

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.floor(x, name=None)` {#floor}

Returns element-wise largest integer not greater than x.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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

*NOTE*: `Maximum` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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

*NOTE*: `Minimum` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.cos(x, name=None)` {#cos}

Computes cos of x element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.sin(x, name=None)` {#sin}

Computes sin of x element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.lbeta(x, name='lbeta')` {#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:


*  <b>`x`</b>: A rank `n + 1` `Tensor` with type `float`, or `double`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  The logarithm of `|Beta(x)|` reducing along the last dimension.

##### Raises:


*  <b>`ValueError`</b>: If `x` is empty with rank one or less.


- - -

### `tf.tan(x, name=None)` {#tan}

Computes tan of x element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.acos(x, name=None)` {#acos}

Computes acos of x element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.asin(x, name=None)` {#asin}

Computes asin of x element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.atan(x, name=None)` {#atan}

Computes atan of x element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.lgamma(x, name=None)` {#lgamma}

Computes the log of the absolute value of `Gamma(x)` element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.digamma(x, name=None)` {#digamma}

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

`Gamma(x)`), element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.erf(x, name=None)` {#erf}

Computes the Gauss error function of `x` element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor` of `SparseTensor`. Must be one of the following types: `half`,
    `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` or `SparseTensor`, respectively. Has the same type as `x`.


- - -

### `tf.erfc(x, name=None)` {#erfc}

Computes the complementary error function of `x` element-wise.

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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

*NOTE*: `SquaredDifference` supports broadcasting. More about broadcasting
[here](http://docs.scipy.org/doc/numpy/user/basics.broadcasting.html)

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `half`, `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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(a) = 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:


*  <b>`a`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`x`</b>: A `Tensor`. Must have the same type as `a`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `a`.


- - -

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

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(a) = 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:


*  <b>`a`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`x`</b>: A `Tensor`. Must have the same type as `a`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `a`.


- - -

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

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:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`q`</b>: A `Tensor`. Must have the same type as `x`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

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

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:


*  <b>`a`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`x`</b>: A `Tensor`. Must have the same type as `a`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `a`.


- - -

### `tf.betainc(a, b, x, name=None)` {#betainc}

Compute the regularized incomplete beta integral \\(I_x(a, b)\\).

The regularized incomplete beta integral is defined as:

```
I_x(a, b) = \frac{B(x; a, b)}{B(a, b)}
```
where

```
B(x; a, b) = \int_0^x t^{a-1} (1 - t)^{b-1} dt
```

is the incomplete beta function and \\(B(a, b)\\) is the *complete*
beta function.

##### Args:


*  <b>`a`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`b`</b>: A `Tensor`. Must have the same type as `a`.
*  <b>`x`</b>: A `Tensor`. Must have the same type as `a`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `a`.


- - -

### `tf.rint(x, name=None)` {#rint}

Returns element-wise integer closest to x.

If the result is midway between two representable values,
the even representable is chosen.
For example:

```
rint(-1.5) ==> -2.0
rint(0.5000001) ==> 1.0
rint([-1.7, -1.5, -0.2, 0.2, 1.5, 1.7, 2.0]) ==> [-2., -2., -0., 0., 2., 2., 2.]
```

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.



## Matrix Math Functions

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

- - -

### `tf.diag(diagonal, name=None)` {#diag}

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:

```prettyprint
# '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:


*  <b>`diagonal`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
    Rank k tensor where k is at most 3.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `diagonal`.


- - -

### `tf.diag_part(input, name=None)` {#diag_part}

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:

```prettyprint
# '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:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `complex64`, `complex128`.
    Rank k tensor where k is 2, 4, or 6.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`. The extracted diagonal.


- - -

### `tf.trace(x, name=None)` {#trace}

Compute the trace of a tensor `x`.

`trace(x)` returns the sum along the main diagonal of each inner-most matrix
in x. If x is of rank `k` with shape `[I, J, K, ..., L, M, N]`, then output
is a tensor of rank `k-2` with dimensions `[I, J, K, ..., L]` where

`output[i, j, k, ..., l] = trace(x[i, j, i, ..., l, :, :])`

For example:

```python
# 'x' is [[1, 2],
#         [3, 4]]
tf.trace(x) ==> 5

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

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

##### Args:


*  <b>`x`</b>: tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  The trace of input tensor.


- - -

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

```python
# '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:


*  <b>`a`</b>: A `Tensor`.
*  <b>`perm`</b>: A permutation of the dimensions of `a`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A transposed `Tensor`.



- - -

### `tf.eye(num_rows, num_columns=None, batch_shape=None, dtype=tf.float32, name=None)` {#eye}

Construct an identity matrix, or a batch of matrices.

```python
# Construct one identity matrix.
tf.eye(2)
==> [[1., 0.],
     [0., 1.]]

# Construct a batch of 3 identity matricies, each 2 x 2.
# batch_identity[i, :, :] is a 2 x 2 identity matrix, i = 0, 1, 2.
batch_identity = tf.eye(2, batch_shape=[3])

# Construct one 2 x 3 "identity" matrix
tf.eye(2, num_columns=3)
==> [[ 1.,  0.,  0.],
     [ 0.,  1.,  0.]]
```

##### Args:


*  <b>`num_rows`</b>: Non-negative `int32` scalar `Tensor` giving the number of rows
    in each batch matrix.
*  <b>`num_columns`</b>: Optional non-negative `int32` scalar `Tensor` giving the number
    of columns in each batch matrix.  Defaults to `num_rows`.
*  <b>`batch_shape`</b>: `int32` `Tensor`.  If provided, returned `Tensor` will have
    leading batch dimensions of this shape.
*  <b>`dtype`</b>: The type of an element in the resulting `Tensor`
*  <b>`name`</b>: A name for this `Op`.  Defaults to "eye".

##### Returns:

  A `Tensor` of shape `batch_shape + [num_rows, num_columns]`


- - -

### `tf.matrix_diag(diagonal, name=None)` {#matrix_diag}

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:

```prettyprint
# '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:


*  <b>`diagonal`</b>: A `Tensor`. Rank `k`, where `k >= 1`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `diagonal`.
  Rank `k+1`, with `output.shape = diagonal.shape + [diagonal.shape[-1]]`.


- - -

### `tf.matrix_diag_part(input, name=None)` {#matrix_diag_part}

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:

```prettyprint
# '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:


*  <b>`input`</b>: A `Tensor`. Rank `k` tensor where `k >= 2`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`.
  The extracted diagonal(s) having shape
  `diagonal.shape = input.shape[:-2] + [min(input.shape[-2:])]`.


- - -

### `tf.matrix_band_part(input, num_lower, num_upper, name=None)` {#matrix_band_part}

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:

```prettyprint
# 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:

```prettyprint
 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:


*  <b>`input`</b>: A `Tensor`. Rank `k` tensor.
*  <b>`num_lower`</b>: A `Tensor` of type `int64`.
    0-D tensor. Number of subdiagonals to keep. If negative, keep entire
    lower triangle.
*  <b>`num_upper`</b>: A `Tensor` of type `int64`.
    0-D tensor. Number of superdiagonals to keep. If negative, keep
    entire upper triangle.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`.
  Rank `k` tensor of the same shape as input. The extracted banded tensor.


- - -

### `tf.matrix_set_diag(input, diagonal, name=None)` {#matrix_set_diag}

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:


*  <b>`input`</b>: A `Tensor`. Rank `k+1`, where `k >= 1`.
*  <b>`diagonal`</b>: A `Tensor`. Must have the same type as `input`.
    Rank `k`, where `k >= 1`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`.
  Rank `k+1`, with `output.shape = input.shape`.


- - -

### `tf.matrix_transpose(a, name='matrix_transpose')` {#matrix_transpose}

Transposes last two dimensions of tensor `a`.

For example:

```python
# Matrix with no batch dimension.
# 'x' is [[1 2 3]
#         [4 5 6]]
tf.matrix_transpose(x) ==> [[1 4]
                                 [2 5]
                                 [3 6]]

# Matrix with two batch dimensions.
# x.shape is [1, 2, 3, 4]
# tf.matrix_transpose(x) is shape [1, 2, 4, 3]
```

##### Args:


*  <b>`a`</b>: A `Tensor` with `rank >= 2`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A transposed batch matrix `Tensor`.

##### Raises:


*  <b>`ValueError`</b>: If `a` is determined statically to have `rank < 2`.



- - -

### `tf.matmul(a, b, transpose_a=False, transpose_b=False, adjoint_a=False, adjoint_b=False, a_is_sparse=False, b_is_sparse=False, name=None)` {#matmul}

Multiplies matrix `a` by matrix `b`, producing `a` * `b`.

The inputs must be matrices (or tensors of rank > 2, representing batches of
matrices), with matching inner dimensions, possibly after transposition.

Both matrices must be of the same type. The supported types are:
`float16`, `float32`, `float64`, `int32`, `complex64`, `complex128`.

Either matrix can be transposed or adjointed (conjugated and transposed) on
the fly by setting one of the corresponding flag to `True`. These are `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.
This optimization is only available for plain matrices (rank-2 tensors) with
datatypes `bfloat16` or `float32`.

For example:

```python
# 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]]


# 3-D tensor `a`
a = tf.constant(np.arange(1, 13, dtype=np.int32),
                shape=[2, 2, 3])                  => [[[ 1.  2.  3.]
                                                       [ 4.  5.  6.]],
                                                      [[ 7.  8.  9.]
                                                       [10. 11. 12.]]]

# 3-D tensor `b`
b = tf.constant(np.arange(13, 25, dtype=np.int32),
                shape=[2, 3, 2])                   => [[[13. 14.]
                                                        [15. 16.]
                                                        [17. 18.]],
                                                       [[19. 20.]
                                                        [21. 22.]
                                                        [23. 24.]]]
c = tf.matmul(a, b) => [[[ 94 100]
                         [229 244]],
                        [[508 532]
                         [697 730]]]
```

##### Args:


*  <b>`a`</b>: `Tensor` of type `float16`, `float32`, `float64`, `int32`, `complex64`,
    `complex128` and rank > 1.
*  <b>`b`</b>: `Tensor` with same type and rank as `a`.
*  <b>`transpose_a`</b>: If `True`, `a` is transposed before multiplication.
*  <b>`transpose_b`</b>: If `True`, `b` is transposed before multiplication.
*  <b>`adjoint_a`</b>: If `True`, `a` is conjugated and transposed before
    multiplication.
*  <b>`adjoint_b`</b>: If `True`, `b` is conjugated and transposed before
    multiplication.
*  <b>`a_is_sparse`</b>: If `True`, `a` is treated as a sparse matrix.
*  <b>`b_is_sparse`</b>: If `True`, `b` is treated as a sparse matrix.
*  <b>`name`</b>: Name for the operation (optional).

##### Returns:

  A `Tensor` of the same type as `a` and `b` where each inner-most matrix is
  the product of the corresponding matrices in `a` and `b`, e.g. if all
  transpose or adjoint attributes are `False`:

  `output`[..., i, j] = sum_k (`a`[..., i, k] * `b`[..., k, j]),
  for all indices i, j.


*  <b>`Note`</b>: This is matrix product, not element-wise product.


##### Raises:


*  <b>`ValueError`</b>: If transpose_a and adjoint_a, or transpose_b and adjoint_b
    are both set to True.



- - -

### `tf.norm(tensor, ord='euclidean', axis=None, keep_dims=False, name=None)` {#norm}

Computes the norm of vectors, matrices, and tensors.

This function can compute 3 different matrix norms (Frobenius, 1-norm, and
inf-norm) and up to 9218868437227405311 different vectors norms.

##### Args:


*  <b>`tensor`</b>: `Tensor` of types `float32`, `float64`, `complex64`, `complex128`
*  <b>`ord`</b>: Order of the norm. Supported values are 'fro', 'euclidean', `0`,
    `1, `2`, `np.inf` and any positive real number yielding the corresponding
    p-norm. Default is 'euclidean' which is equivalent to Frobenius norm if
    `tensor` is a matrix and equivalent to 2-norm for vectors.
    Some restrictions apply,
      a) The Frobenius norm `fro` is not defined for vectors,
      b) If axis is a 2-tuple (matrix-norm), only 'euclidean', 'fro', `1`,
         `np.inf` are supported.
    See the description of `axis` on how to compute norms for a batch of
    vectors or matrices stored in a tensor.
*  <b>`axis`</b>: If `axis` is `None` (the default), the input is considered a vector
    and a single vector norm is computed over the entire set of values in the
    tensor, i.e. `norm(tensor, ord=ord)` is equivalent to
    `norm(reshape(tensor, [-1]), ord=ord)`.
    If `axis` is a Python integer, the input is considered a batch of vectors,
    and `axis`t determines the axis in `tensor` over which to compute vector
    norms.
    If `axis` is a 2-tuple of Python integers it is considered a batch of
    matrices and `axis` determines the axes in `tensor` over which to compute
    a matrix norm.
    Negative indices are supported. Example: If you are passing a tensor that
    can be either a matrix or a batch of matrices at runtime, pass
    `axis=[-2,-1]` instead of `axis=None` to make sure that matrix norms are
    computed.
*  <b>`keep_dims`</b>: If True, the axis indicated in `axis` are kept with size 1.
    Otherwise, the dimensions in `axis` are removed from the output shape.
*  <b>`name`</b>: The name of the op.

##### Returns:


*  <b>`output`</b>: A `Tensor` of the same type as tensor, containing the vector or
    matrix norms. If `keep_dims` is True then the rank of output is equal to
    the rank of `tensor`. Otherwise, if `axis` is none the output is a scalar,
    if `axis` is an integer, the rank of `output` is one less than the rank
    of `tensor`, if `axis` is a 2-tuple the rank of `output` is two less
    than the rank of `tensor`.

##### Raises:


*  <b>`ValueError`</b>: If `ord` or `axis` is invalid.

@compatibility(numpy)
Mostly equivalent to numpy.linalg.norm.
Not supported: ord <= 0, 2-norm for matrices, nuclear norm.

##### Other differences:

  a) If axis is `None`, treats the the flattened `tensor` as a vector
   regardless of rank.
  b) Explicitly supports 'euclidean' norm as the default, including for
   higher order tensors.
@end_compatibility


- - -

### `tf.matrix_determinant(input, name=None)` {#matrix_determinant}

Computes the determinant of one ore more square matrices.

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

##### Args:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
    Shape is `[..., M, M]`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`. Shape is `[...]`.


- - -

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

Computes the inverse of one or more 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:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float64`, `float32`.
    Shape is `[..., M, M]`.
*  <b>`adjoint`</b>: An optional `bool`. Defaults to `False`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`. Shape is `[..., M, M]`.

  @compatibility(numpy)
  Equivalent to np.linalg.inv
  @end_compatibility


- - -

### `tf.cholesky(input, name=None)` {#cholesky}

Computes the Cholesky decomposition of one or more 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:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float64`, `float32`.
    Shape is `[..., M, M]`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `input`. Shape is `[..., M, M]`.


- - -

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

Solves systems of linear eqns `A X = RHS`, given Cholesky factorizations.

```python
# Solve 10 separate 2x2 linear systems:
A = ... # shape 10 x 2 x 2
RHS = ... # shape 10 x 2 x 1
chol = tf.cholesky(A)  # shape 10 x 2 x 2
X = tf.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:


*  <b>`chol`</b>: A `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 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.
*  <b>`rhs`</b>: A `Tensor`, same type as `chol`, shape is `[..., M, K]`.
*  <b>`name`</b>: A name to give this `Op`.  Defaults to `cholesky_solve`.

##### Returns:

  Solution to `A x = rhs`, shape `[..., M, K]`.


- - -

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

Solves systems of linear equations.

`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:


*  <b>`matrix`</b>: A `Tensor`. Must be one of the following types: `float64`, `float32`, `complex64`, `complex128`.
    Shape is `[..., M, M]`.
*  <b>`rhs`</b>: A `Tensor`. Must have the same type as `matrix`.
    Shape is `[..., M, K]`.
*  <b>`adjoint`</b>: An optional `bool`. Defaults to `False`.
    Boolean indicating whether to solve with `matrix` or its (block-wise)
    adjoint.
*  <b>`name`</b>: A name for the operation (optional).

##### 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)` {#matrix_triangular_solve}

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[..., :, :]`.
If `adjoint` is `False` then the strictly then the  innermost matrices in
`output` satisfy matrix equations
`adjoint(matrix[..., i, k]) * output[..., k, j] = rhs[..., i, j]`.

##### Args:


*  <b>`matrix`</b>: A `Tensor`. Must be one of the following types: `float64`, `float32`.
    Shape is `[..., M, M]`.
*  <b>`rhs`</b>: A `Tensor`. Must have the same type as `matrix`.
    Shape is `[..., M, K]`.
*  <b>`lower`</b>: An optional `bool`. Defaults to `True`.
    Boolean indicating whether the innermost matrices in `matrix` are
    lower or upper triangular.
*  <b>`adjoint`</b>: An optional `bool`. Defaults to `False`.
    Boolean indicating whether to solve with `matrix` or its (block-wise)
             adjoint.

    @compatibility(numpy)
    Equivalent to np.linalg.triangular_solve
    @end_compatibility

*  <b>`name`</b>: A name for the operation (optional).

##### 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)` {#matrix_solve_ls}

Solves one or more 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:


*  <b>`matrix`</b>: `Tensor` of shape `[..., M, N]`.
*  <b>`rhs`</b>: `Tensor` of shape `[..., M, K]`.
*  <b>`l2_regularizer`</b>: 0-D `double` `Tensor`. Ignored if `fast=False`.
*  <b>`fast`</b>: bool. Defaults to `True`.
*  <b>`name`</b>: string, optional name of the operation.

##### Returns:


*  <b>`output`</b>: `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.


- - -

### `tf.qr(input, full_matrices=None, name=None)` {#qr}

Computes the QR decompositions of one or more matrices.

Computes the QR decomposition of each inner matrix in `tensor` such that
`tensor[..., :, :] = q[..., :, :] * r[..., :,:])`

```prettyprint
# a is a tensor.
# q is a tensor of orthonormal matrices.
# r is a tensor of upper triangular matrices.
q, r = qr(a)
q_full, r_full = qr(a, full_matrices=True)
```

##### Args:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float64`, `float32`, `complex64`, `complex128`.
    A tensor of shape `[..., M, N]` whose inner-most 2 dimensions
    form matrices of size `[M, N]`. Let `P` be the minimum of `M` and `N`.
*  <b>`full_matrices`</b>: An optional `bool`. Defaults to `False`.
    If true, compute full-sized `q` and `r`. If false
    (the default), compute only the leading `P` columns of `q`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A tuple of `Tensor` objects (q, r).

*  <b>`q`</b>: A `Tensor`. Has the same type as `input`. Orthonormal basis for range of `a`. If `full_matrices` is `False` then
    shape is `[..., M, P]`; if `full_matrices` is `True` then shape is
    `[..., M, M]`.
*  <b>`r`</b>: A `Tensor`. Has the same type as `input`. Triangular factor. If `full_matrices` is `False` then shape is
    `[..., P, N]`. If `full_matrices` is `True` then shape is `[..., M, N]`.


- - -

### `tf.self_adjoint_eig(tensor, name=None)` {#self_adjoint_eig}

Computes the eigen decomposition of a batch of self-adjoint matrices.

Computes the eigenvalues and eigenvectors of the innermost N-by-N matrices
in `tensor` such that
`tensor[...,:,:] * v[..., :,i] = e[..., i] * v[...,:,i]`, for i=0...N-1.

##### Args:


*  <b>`tensor`</b>: `Tensor` of shape `[..., N, N]`. Only the lower triangular part of
    each inner inner matrix is referenced.
*  <b>`name`</b>: string, optional name of the operation.

##### Returns:


*  <b>`e`</b>: Eigenvalues. Shape is `[..., N]`.
*  <b>`v`</b>: Eigenvectors. Shape is `[..., N, N]`. The columns of the inner most
    matrices contain eigenvectors of the corresponding matrices in `tensor`


- - -

### `tf.self_adjoint_eigvals(tensor, name=None)` {#self_adjoint_eigvals}

Computes the eigenvalues of one or more self-adjoint matrices.

##### Args:


*  <b>`tensor`</b>: `Tensor` of shape `[..., N, N]`.
*  <b>`name`</b>: string, optional name of the operation.

##### Returns:


*  <b>`e`</b>: Eigenvalues. Shape is `[..., N]`. The vector `e[..., :]` contains the `N`
    eigenvalues of `tensor[..., :, :]`.


- - -

### `tf.svd(tensor, full_matrices=False, compute_uv=True, name=None)` {#svd}

Computes the singular value decompositions of one or more matrices.

Computes the SVD of each inner matrix in `tensor` such that
`tensor[..., :, :] = u[..., :, :] * diag(s[..., :, :]) * transpose(v[..., :,
:])`

```prettyprint
# a is a tensor.
# s is a tensor of singular values.
# u is a tensor of left singular vectors.
#v is a tensor of right singular vectors.
s, u, v = svd(a)
s = svd(a, compute_uv=False)
```

##### Args:


*  <b>`tensor`</b>: `Tensor` of shape `[..., M, N]`. Let `P` be the minimum of `M` and
    `N`.
*  <b>`full_matrices`</b>: If true, compute full-sized `u` and `v`. If false
    (the default), compute only the leading `P` singular vectors.
    Ignored if `compute_uv` is `False`.
*  <b>`compute_uv`</b>: If `True` then left and right singular vectors will be
    computed and returned in `u` and `v`, respectively. Otherwise, only the
    singular values will be computed, which can be significantly faster.
*  <b>`name`</b>: string, optional name of the operation.

##### Returns:


*  <b>`s`</b>: Singular values. Shape is `[..., P]`.
*  <b>`u`</b>: Right singular vectors. If `full_matrices` is `False` (default) then
    shape is `[..., M, P]`; if `full_matrices` is `True` then shape is
    `[..., M, M]`. Not returned if `compute_uv` is `False`.
*  <b>`v`</b>: Left singular vectors. If `full_matrices` is `False` (default) then
    shape is `[..., N, P]`. If `full_matrices` is `True` then shape is
    `[..., N, N]`. Not returned if `compute_uv` is `False`.

@compatibility(numpy)
Mostly equivalent to numpy.linalg.svd, except that the order of output
arguments here is `s`, `u`, `v` when `compute_uv` is `True`, as opposed to
`u`, `s`, `v` for numpy.linalg.svd.
@end_compatibility




## Tensor Math Function

TensorFlow provides operations that you can use to add tensor functions to your
graph.

- - -

### `tf.tensordot(a, b, axes, name=None)` {#tensordot}

Tensor contraction of a and b along specified axes.

Tensordot (also known as tensor contraction) sums the product of elements
from `a` and `b` over the indices specified by `a_axes` and `b_axes`.
The lists `a_axes` and `b_axes` specify those pairs of axes along which to
contract the tensors. The axis `a_axes[i]` of `a` must have the same dimension
as axis `b_axes[i]` of `b` for all `i` in `range(0, len(a_axes))`. The lists
`a_axes` and `b_axes` must have identical length and consist of unique
integers that specify valid axes for each of the tensors.

This operation corresponds to `numpy.tensordot(a, b, axes)`.

Example 1: When `a` and `b` are matrices (order 2), the case `axes = 1`
is equivalent to matrix multiplication.

Example 2: When `a` and `b` are matrices (order 2), the case
`axes = [[1], [0]]` is equivalent to matrix multiplication.

Example 3: Suppose that \\(a_ijk\\) and \\(b_lmn\\) represent two
tensors of order 3. Then, `contract(a, b, [0], [2])` is the order 4 tensor
\\(c_{jklm}\\) whose entry
corresponding to the indices \\((j,k,l,m)\\) is given by:

\\( c_{jklm} = \sum_i a_{ijk} b_{lmi} \\).

In general, `order(c) = order(a) + order(b) - 2*len(axes[0])`.

##### Args:


*  <b>`a`</b>: `Tensor` of type `float32` or `float64`.
*  <b>`b`</b>: `Tensor` with the same type as `a`.
*  <b>`axes`</b>: Either a scalar `N`, or a list or an `int32` `Tensor` of shape [2, k].
   If axes is a scalar, sum over the last N axes of a and the first N axes
   of b in order.
   If axes is a list or `Tensor` the first and second row contain the set of
   unique integers specifying axes along which the contraction is computed,
   for `a` and `b`, respectively. The number of axes for `a` and `b` must
   be equal.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` with the same type as `a`.

##### Raises:


*  <b>`ValueError`</b>: If the shapes of `a`, `b`, and `axes` are incompatible.
*  <b>`IndexError`</b>: If the values in axes exceed the rank of the corresponding
    tensor.




## 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)` {#complex}

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:


*  <b>`real`</b>: A `Tensor`. Must be one of the following types: `float32`,
    `float64`.
*  <b>`imag`</b>: A `Tensor`. Must have the same type as `real`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of type `complex64` or `complex128`.


- - -

### `tf.conj(x, name=None)` {#conj}

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]

If `x` is real, it is returned unchanged.

##### Args:


*  <b>`x`</b>: `Tensor` to conjugate.  Must have numeric type.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` that is the conjugate of `x` (with the same type).

##### Raises:


*  <b>`TypeError`</b>: If `x` is not a numeric tensor.


- - -

### `tf.imag(input, name=None)` {#imag}

Returns the imaginary part of a complex number.

Given a tensor `input` of complex numbers, this operation returns a tensor of
type `float32` or `float64` 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:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `complex64`,
    `complex128`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of type `float32` or `float64`.


- - -

### `tf.real(input, name=None)` {#real}

Returns the real part of a complex number.

Given a tensor `input` of complex numbers, this operation returns a tensor of
type `float32` or `float64` 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]
```

If `input` is already real, it is returned unchanged.

##### Args:


*  <b>`input`</b>: A `Tensor`. Must have numeric type.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of type `float32` or `float64`.



## Fourier Transform Functions

TensorFlow provides several operations that you can use to add discrete
Fourier transform functions to your graph.

- - -

### `tf.fft(input, name=None)` {#fft}

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

dimension of `input`.

##### Args:


*  <b>`input`</b>: A `Tensor` of type `complex64`. A complex64 tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.ifft(input, name=None)` {#ifft}

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

dimension of `input`.

##### Args:


*  <b>`input`</b>: A `Tensor` of type `complex64`. A complex64 tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.fft2d(input, name=None)` {#fft2d}

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

2 dimensions of `input`.

##### Args:


*  <b>`input`</b>: A `Tensor` of type `complex64`. A complex64 tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.

  @compatibility(numpy)
  Equivalent to np.fft2
  @end_compatibility


- - -

### `tf.ifft2d(input, name=None)` {#ifft2d}

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

2 dimensions of `input`.

##### Args:


*  <b>`input`</b>: A `Tensor` of type `complex64`. A complex64 tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.

  @compatibility(numpy)
  Equivalent to np.ifft2
  @end_compatibility


- - -

### `tf.fft3d(input, name=None)` {#fft3d}

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

dimensions of `input`.

##### Args:


*  <b>`input`</b>: A `Tensor` of type `complex64`. A complex64 tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.

  @compatibility(numpy)
  Equivalent to np.fft3
  @end_compatibility


- - -

### `tf.ifft3d(input, name=None)` {#ifft3d}

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

3 dimensions of `input`.

##### Args:


*  <b>`input`</b>: A `Tensor` of type `complex64`. A complex64 tensor.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.

  @compatibility(numpy)
  Equivalent to np.fft3
  @end_compatibility



## 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, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_sum}

Computes the sum of elements across dimensions of a tensor.

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

For example:

```python
# '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:


*  <b>`input_tensor`</b>: The tensor to reduce. Should have numeric type.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.sum
@end_compatibility


- - -

### `tf.reduce_prod(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_prod}

Computes the product of elements across dimensions of a tensor.

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

##### Args:


*  <b>`input_tensor`</b>: The tensor to reduce. Should have numeric type.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.prod
@end_compatibility


- - -

### `tf.reduce_min(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_min}

Computes the minimum of elements across dimensions of a tensor.

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

##### Args:


*  <b>`input_tensor`</b>: The tensor to reduce. Should have numeric type.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.min
@end_compatibility


- - -

### `tf.reduce_max(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_max}

Computes the maximum of elements across dimensions of a tensor.

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

##### Args:


*  <b>`input_tensor`</b>: The tensor to reduce. Should have numeric type.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.max
@end_compatibility


- - -

### `tf.reduce_mean(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_mean}

Computes the mean of elements across dimensions of a tensor.

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

For example:

```python
# '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:


*  <b>`input_tensor`</b>: The tensor to reduce. Should have numeric type.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.mean
@end_compatibility


- - -

### `tf.reduce_all(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_all}

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

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

For example:

```python
# '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:


*  <b>`input_tensor`</b>: The boolean tensor to reduce.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.all
@end_compatibility


- - -

### `tf.reduce_any(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_any}

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

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

For example:

```python
# '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:


*  <b>`input_tensor`</b>: The boolean tensor to reduce.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.

@compatibility(numpy)
Equivalent to np.any
@end_compatibility


- - -

### `tf.reduce_logsumexp(input_tensor, axis=None, keep_dims=False, name=None, reduction_indices=None)` {#reduce_logsumexp}

Computes log(sum(exp(elements across dimensions of a tensor))).

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

This function is more numerically stable than log(sum(exp(input))). It avoids
overflows caused by taking the exp of large inputs and underflows caused by
taking the log of small inputs.

For example:

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

##### Args:


*  <b>`input_tensor`</b>: The tensor to reduce. Should have numeric type.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor.


- - -

### `tf.count_nonzero(input_tensor, axis=None, keep_dims=False, dtype=tf.int64, name=None, reduction_indices=None)` {#count_nonzero}

Computes number of nonzero elements across dimensions of a tensor.

Reduces `input_tensor` along the dimensions given in `axis`.
Unless `keep_dims` is true, the rank of the tensor is reduced by 1 for each
entry in `axis`. If `keep_dims` is true, the reduced dimensions
are retained with length 1.

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

**NOTE** Floating point comparison to zero is done by exact floating point
equality check.  Small values are **not** rounded to zero for purposes of
the nonzero check.

For example:

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

##### Args:


*  <b>`input_tensor`</b>: The tensor to reduce. Should be of numeric type, or `bool`.
*  <b>`axis`</b>: The dimensions to reduce. If `None` (the default),
    reduces all dimensions.
*  <b>`keep_dims`</b>: If true, retains reduced dimensions with length 1.
*  <b>`dtype`</b>: The output dtype; defaults to `tf.int64`.
*  <b>`name`</b>: A name for the operation (optional).
*  <b>`reduction_indices`</b>: The old (deprecated) name for axis.

##### Returns:

  The reduced tensor (number of nonzero values).



- - -

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

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.

NOTE: This operation is not differentiable and cannot be used if inputs depend
on trainable variables. Please use `tf.add_n` for such cases.

For example:

```python
# 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:


*  <b>`inputs`</b>: A list of `Tensor` objects, each with same shape and type.
*  <b>`shape`</b>: Shape of elements of `inputs`.
*  <b>`tensor_dtype`</b>: The type of `inputs`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of same shape and type as the elements of `inputs`.

##### Raises:


*  <b>`ValueError`</b>: If `inputs` don't all have same shape and dtype or the shape
  cannot be inferred.



- - -

### `tf.einsum(equation, *inputs)` {#einsum}

A generalized contraction between tensors of arbitrary dimension.

This function returns a tensor whose elements are defined by `equation`,
which is written in a shorthand form inspired by the Einstein summation
convention.  As an example, consider multiplying two matrices
A and B to form a matrix C.  The elements of C are given by:

```
  C[i,k] = sum_j A[i,j] * B[j,k]
```

The corresponding `equation` is:

```
  ij,jk->ik
```

In general, the `equation` is obtained from the more familiar element-wise
equation by
  1. removing variable names, brackets, and commas,
  2. replacing "*" with ",",
  3. dropping summation signs, and
  4. moving the output to the right, and replacing "=" with "->".

Many common operations can be expressed in this way.  For example:

```python
# Matrix multiplication
>>> einsum('ij,jk->ik', m0, m1)  # output[i,k] = sum_j m0[i,j] * m1[j, k]

# Dot product
>>> einsum('i,i->', u, v)  # output = sum_i u[i]*v[i]

# Outer product
>>> einsum('i,j->ij', u, v)  # output[i,j] = u[i]*v[j]

# Transpose
>>> einsum('ij->ji', m)  # output[j,i] = m[i,j]

# Batch matrix multiplication
>>> einsum('aij,ajk->aik', s, t)  # out[a,i,k] = sum_j s[a,i,j] * t[a, j, k]
```

This function behaves like `numpy.einsum`, but does not support:
* Ellipses (subscripts like `ij...,jk...->ik...`)
* Subscripts where an axis appears more than once for a single input
  (e.g. `ijj,k->ik`).
* Subscripts that are summed across multiple inputs (e.g., `ij,ij,jk->ik`).

##### Args:


*  <b>`equation`</b>: a `str` describing the contraction, in the same format as
    `numpy.einsum`.
*  <b>`inputs`</b>: the inputs to contract (each one a `Tensor`), whose shapes should
    be consistent with `equation`.

##### Returns:

  The contracted `Tensor`, with shape determined by `equation`.

##### Raises:


*  <b>`ValueError`</b>: If
    - the format of `equation` is incorrect,
    - the number of inputs implied by `equation` does not match `len(inputs)`,
    - an axis appears in the output subscripts but not in any of the inputs,
    - the number of dimensions of an input differs from the number of
      indices in its subscript, or
    - the input shapes are inconsistent along a particular axis.



## Scan

TensorFlow provides several operations that you can use to perform scans
(running totals) across one axis of a tensor.

- - -

### `tf.cumsum(x, axis=0, exclusive=False, reverse=False, name=None)` {#cumsum}

Compute the cumulative sum of the tensor `x` along `axis`.

By default, this op performs an inclusive cumsum, which means that the first
element of the input is identical to the first element of the output:
```prettyprint
tf.cumsum([a, b, c]) ==> [a, a + b, a + b + c]
```

By setting the `exclusive` kwarg to `True`, an exclusive cumsum is performed
instead:
```prettyprint
tf.cumsum([a, b, c], exclusive=True) ==> [0, a, a + b]
```

By setting the `reverse` kwarg to `True`, the cumsum is performed in the
opposite direction:
```prettyprint
tf.cumsum([a, b, c], reverse=True) ==> [a + b + c, b + c, c]
```
This is more efficient than using separate `tf.reverse` ops.

The `reverse` and `exclusive` kwargs can also be combined:
```prettyprint
tf.cumsum([a, b, c], exclusive=True, reverse=True) ==> [b + c, c, 0]
```

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`,
     `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`,
     `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`axis`</b>: A `Tensor` of type `int32` (default: 0).
*  <b>`reverse`</b>: A `bool` (default: False).
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.


- - -

### `tf.cumprod(x, axis=0, exclusive=False, reverse=False, name=None)` {#cumprod}

Compute the cumulative product of the tensor `x` along `axis`.

By default, this op performs an inclusive cumprod, which means that the
first
element of the input is identical to the first element of the output:
```prettyprint
tf.cumprod([a, b, c]) ==> [a, a * b, a * b * c]
```

By setting the `exclusive` kwarg to `True`, an exclusive cumprod is
performed
instead:
```prettyprint
tf.cumprod([a, b, c], exclusive=True) ==> [1, a, a * b]
```

By setting the `reverse` kwarg to `True`, the cumprod is performed in the
opposite direction:
```prettyprint
tf.cumprod([a, b, c], reverse=True) ==> [a * b * c, b * c, c]
```
This is more efficient than using separate `tf.reverse` ops.

The `reverse` and `exclusive` kwargs can also be combined:
```prettyprint
tf.cumprod([a, b, c], exclusive=True, reverse=True) ==> [b * c, c, 1]
```

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`,
     `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`,
     `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`axis`</b>: A `Tensor` of type `int32` (default: 0).
*  <b>`reverse`</b>: A `bool` (default: False).
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`.



## 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:

```python
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)` {#segment_sum}

Computes the sum along segments of a tensor.

Read [the section on Segmentation](../../api_docs/python/math_ops.md#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`.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/SegmentSum.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`segment_ids`</b>: A `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.
*  <b>`name`</b>: A name for the operation (optional).

##### 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)` {#segment_prod}

Computes the product along segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#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`.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/SegmentProd.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`segment_ids`</b>: A `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.
*  <b>`name`</b>: A name for the operation (optional).

##### 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)` {#segment_min}

Computes the minimum along segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#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`.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/SegmentMin.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`, `half`.
*  <b>`segment_ids`</b>: A `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.
*  <b>`name`</b>: A name for the operation (optional).

##### 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)` {#segment_max}

Computes the maximum along segments of a tensor.

Read [the section on Segmentation](../../api_docs/python/math_ops.md#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`.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/SegmentMax.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`, `half`.
*  <b>`segment_ids`</b>: A `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.
*  <b>`name`</b>: A name for the operation (optional).

##### 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_mean(data, segment_ids, name=None)` {#segment_mean}

Computes the mean along segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#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.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/SegmentMean.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`, `half`.
*  <b>`segment_ids`</b>: A `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.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.unsorted_segment_sum(data, segment_ids, num_segments, name=None)` {#unsorted_segment_sum}

Computes the sum along segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#segmentation) for an explanation
of segments.

Computes a tensor such that
`(output[i] = sum_{j...} data[j...]` where the sum is over tuples `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.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/UnsortedSegmentSum.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`segment_ids`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`.
    A tensor whose shape is a prefix of `data.shape`.
*  <b>`num_segments`</b>: A `Tensor` of type `int32`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `data`.
  Has same shape as data, except for the first `segment_ids.rank`
  dimensions, which are replaced with a single dimension which has size
  `num_segments`.


- - -

### `tf.unsorted_segment_max(data, segment_ids, num_segments, name=None)` {#unsorted_segment_max}

Computes the Max along segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#segmentation) for an explanation
of segments.

This operator is similar to the [unsorted segment sum operator](../../api_docs/python/math_ops.md#UnsortedSegmentSum).
Instead of computing the sum over segments, it computes the maximum
such that:

\\(output_i = \max_j data_j\\) where max is over `j` such
that `segment_ids[j] == i`.

If the maximum is empty for a given segment ID `i`, it outputs the smallest possible value for specific numeric type,
 `output[i] = numeric_limits<T>::min()`.

<div style="width:70%; margin:auto; margin-bottom:10px; margin-top:20px;">
<img style="width:100%" src="../../images/UnsortedSegmentSum.png" alt>
</div>

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`, `half`.
*  <b>`segment_ids`</b>: A `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.
*  <b>`num_segments`</b>: A `Tensor` of type `int32`.
*  <b>`name`</b>: A name for the operation (optional).

##### 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)` {#sparse_segment_sum}

Computes the sum along sparse segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#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:

```prettyprint
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:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int32`, `int64`, `uint8`, `int16`, `int8`, `uint16`, `half`.
*  <b>`indices`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`.
    A 1-D tensor. Has same rank as `segment_ids`.
*  <b>`segment_ids`</b>: A `Tensor` of type `int32`.
    A 1-D tensor. Values should be sorted and can be repeated.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.sparse_segment_mean(data, indices, segment_ids, name=None)` {#sparse_segment_mean}

Computes the mean along sparse segments of a tensor.

Read [the section on
Segmentation](../../api_docs/python/math_ops.md#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:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`indices`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`.
    A 1-D tensor. Has same rank as `segment_ids`.
*  <b>`segment_ids`</b>: A `Tensor` of type `int32`.
    A 1-D tensor. Values should be sorted and can be repeated.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.sparse_segment_sqrt_n(data, indices, segment_ids, name=None)` {#sparse_segment_sqrt_n}

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](../../api_docs/python/math_ops.md#segmentation) for an explanation
of segments.

##### Args:


*  <b>`data`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`.
*  <b>`indices`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`.
    A 1-D tensor. Has same rank as `segment_ids`.
*  <b>`segment_ids`</b>: A `Tensor` of type `int32`.
    A 1-D tensor. Values should be sorted and can be repeated.
*  <b>`name`</b>: A name for the operation (optional).

##### 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.




## 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, axis=None, name=None, dimension=None)` {#argmin}

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

##### Args:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`axis`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`.
    int32, 0 <= axis < rank(input).  Describes which axis
    of the input Tensor to reduce across. For vectors, use axis = 0.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of type `int64`.


- - -

### `tf.argmax(input, axis=None, name=None, dimension=None)` {#argmax}

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

##### Args:


*  <b>`input`</b>: A `Tensor`. Must be one of the following types: `float32`, `float64`, `int64`, `int32`, `uint8`, `uint16`, `int16`, `int8`, `complex64`, `complex128`, `qint8`, `quint8`, `qint32`, `half`.
*  <b>`axis`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`.
    int32, 0 <= axis < rank(input).  Describes which axis
    of the input Tensor to reduce across. For vectors, use axis = 0.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor` of type `int64`.



- - -

### `tf.setdiff1d(x, y, index_dtype=tf.int32, name=None)` {#setdiff1d}

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:

```prettyprint
x = [1, 2, 3, 4, 5, 6]
y = [1, 3, 5]
```

This operation would return:

```prettyprint
out ==> [2, 4, 6]
idx ==> [1, 3, 5]
```

##### Args:


*  <b>`x`</b>: A `Tensor`. 1-D. Values to keep.
*  <b>`y`</b>: A `Tensor`. Must have the same type as `x`. 1-D. Values to remove.
*  <b>`out_idx`</b>: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A tuple of `Tensor` objects (out, idx).

*  <b>`out`</b>: A `Tensor`. Has the same type as `x`. 1-D. Values present in `x` but not in `y`.
*  <b>`idx`</b>: A `Tensor` of type `out_idx`. 1-D. Positions of `x` values preserved in `out`.


- - -

### `tf.where(condition, x=None, y=None, name=None)` {#where}

Return the elements, either from `x` or `y`, depending on the `condition`.

If both `x` and `y` are None, then this operation returns the coordinates of
true elements of `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 input. Indices are output in row-major
order.

If both non-None, `x` and `y` must have the same shape.
The `condition` tensor must be a scalar if `x` and `y` are scalar.
If `x` and `y` are vectors or higher rank, then `condition` must be either a
vector with size matching the first dimension of `x`, or must have the same
shape as `x`.

The `condition` tensor acts as a mask that chooses, based on the value at each
element, whether the corresponding element / row in the output should be taken
from `x` (if true) or `y` (if false).

If `condition` is a vector and `x` and `y` are higher rank matrices, then it
chooses which row (outer dimension) to copy from `x` and `y`. If `condition`
has the same shape as `x` and `y`, then it chooses which element to copy from
`x` and `y`.

##### Args:


*  <b>`condition`</b>: A `Tensor` of type `bool`
*  <b>`x`</b>: A Tensor which may have the same shape as `condition`. If `condition` is
    rank 1, `x` may have higher rank, but its first dimension must match the
    size of `condition`.
*  <b>`y`</b>: A `tensor` with the same shape and type as `x`.
*  <b>`name`</b>: A name of the operation (optional)

##### Returns:

  A `Tensor` with the same type and shape as `x`, `y` if they are non-None.
  A `Tensor` with shape `(num_true, dim_size(condition))`.

##### Raises:


*  <b>`ValueError`</b>: When exactly one of `x` or `y` is non-None.


- - -

### `tf.unique(x, out_idx=None, name=None)` {#unique}

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:

```prettyprint
# 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:


*  <b>`x`</b>: A `Tensor`. 1-D.
*  <b>`out_idx`</b>: An optional `tf.DType` from: `tf.int32, tf.int64`. Defaults to `tf.int32`.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A tuple of `Tensor` objects (y, idx).

*  <b>`y`</b>: A `Tensor`. Has the same type as `x`. 1-D.
*  <b>`idx`</b>: A `Tensor` of type `out_idx`. 1-D.



- - -

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

```python
# '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:

```python
# '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:


*  <b>`hypothesis`</b>: A `SparseTensor` containing hypothesis sequences.
*  <b>`truth`</b>: A `SparseTensor` containing truth sequences.
*  <b>`normalize`</b>: A `bool`. If `True`, normalizes the Levenshtein distance by
    length of `truth.`
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A dense `Tensor` with rank `R - 1`, where R is the rank of the
  `SparseTensor` inputs `hypothesis` and `truth`.

##### Raises:


*  <b>`TypeError`</b>: If either `hypothesis` or `truth` are not a `SparseTensor`.



- - -

### `tf.invert_permutation(x, name=None)` {#invert_permutation}

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:

```prettyprint
# tensor `x` is [3, 4, 0, 2, 1]
invert_permutation(x) ==> [2, 4, 3, 0, 1]
```

##### Args:


*  <b>`x`</b>: A `Tensor`. Must be one of the following types: `int32`, `int64`. 1-D.
*  <b>`name`</b>: A name for the operation (optional).

##### Returns:

  A `Tensor`. Has the same type as `x`. 1-D.