shared - defines aesara.shared#

Using Shared Variables#

It is possible to make a function with an internal state. For example, let’s say we want to make an accumulator: at the beginning, the state is initialized to zero, then, on each function call, the state is incremented by the function’s argument.

First let’s define the accumulator function. It adds its argument to the internal state and returns the old state value.

>>> from aesara import shared
>>> state = shared(0)
>>> inc = at.iscalar('inc')
>>> accumulator = function([inc], state, updates=[(state, state+inc)])

This code introduces a few new concepts. The shared function constructs so-called shared variables. These are hybrid symbolic and non-symbolic variables whose value may be shared between multiple functions. Shared variables can be used in symbolic expressions just like the objects returned by dmatrices but they also have an internal value that defines the value taken by this symbolic variable in all the functions that use it. It is called a shared variable because its value is shared between many functions. The value can be accessed and modified by the get_value() and set_value() methods. We will come back to this soon.

The other new thing in this code is the updates parameter of aesara.function(). updates must be supplied with a list of pairs of the form (shared-variable, new expression). It can also be a dictionary whose keys are shared-variables and values are the new expressions. Either way, it means “whenever this function runs, it will replace the value of each shared variable with the result of the corresponding expression”. Above, our accumulator replaces the state’s value with the sum of the state and the increment amount.

Let’s try it out!

>>> print(state.get_value())
>>> accumulator(1)
>>> print(state.get_value())
>>> accumulator(300)
>>> print(state.get_value())

It is possible to reset the state. Just use the .set_value() method:

>>> state.set_value(-1)
>>> accumulator(3)
>>> print(state.get_value())

As we mentioned above, you can define more than one function to use the same shared variable. These functions can all update the value.

>>> decrementor = function([inc], state, updates=[(state, state-inc)])
>>> decrementor(2)
>>> print(state.get_value())

You might be wondering why the updates mechanism exists. You can always achieve a similar result by returning the new expressions, and working with them in NumPy as usual. The updates mechanism can be a syntactic convenience, but it is mainly there for efficiency. Updates to shared variables can sometimes be done more quickly using in-place algorithms (e.g. low-rank matrix updates).

It may happen that you expressed some formula using a shared variable, but you do not want to use its value. In this case, you can use the givens parameter of aesara.function() which replaces a particular node in a graph for the purpose of one particular function.

>>> fn_of_state = state * 2 + inc
>>> # The type of foo must match the shared variable we are replacing
>>> # with the ``givens``
>>> foo = at.scalar(dtype=state.dtype)
>>> skip_shared = function([inc, foo], fn_of_state, givens=[(state, foo)])
>>> skip_shared(1, 3)  # we're using 3 for the state, not state.value
>>> print(state.get_value())  # old state still there, but we didn't use it

The givens parameter can be used to replace any symbolic variable, not just a shared variable. You can replace constants, and expressions, in general. Be careful though, not to allow the expressions introduced by a givens substitution to be co-dependent, the order of substitution is not defined, so the substitutions have to work in any order.

In practice, a good way of thinking about the givens is as a mechanism that allows you to replace any part of your formula with a different expression that evaluates to a tensor of same shape and dtype.


Aesara shared variable broadcast pattern default to False for each dimensions. Shared variable size can change over time, so we can’t use the shape to find the broadcastable pattern. If you want a different pattern, just pass it as a parameter aesara.shared(..., broadcastable=(True, False))


Use the shape parameter to specify tuples of static shapes instead; the old broadcastable values are being phased-out. Unknown shape values for dimensions take the value None; otherwise, integers are used for known static shape values. For example, aesara.shared(..., shape=(1, None)).


class aesara.compile.sharedvalue.SharedVariable[source]#

Variable with storage that is shared between the compiled functions that it appears in. These variables are meant to be created by registered shared constructors (see shared_constructor()).

The user-friendly constructor is shared()

get_value(self, borrow=False, return_internal_type=False)[source]#
  • borrow (bool) – True to permit returning of an object aliased to internal memory.

  • return_internal_type (bool) – True to permit the returning of an arbitrary type object used internally to store the shared variable.

By default, return a copy of the data. If borrow=True (and return_internal_type=False), maybe it will return a copy. For tensor, it will always return an ndarray by default, so if the data is on another device, it will return a copy, but if the data is on the CPU, it will return the original data. If you do borrow=True and return_internal_type=True, it will always return the original data, not a copy, but this can be a non-ndarray type of object.

set_value(self, new_value, borrow=False)[source]#
  • new_value (A compatible type for this shared variable.) – The new value.

  • borrow (bool) – True to use the new_value directly, potentially creating problems related to aliased memory.

The new value will be seen by all functions using this SharedVariable.

__init__(self, name, type, value, strict, container=None)[source]#
  • name (None or str) – The name for this variable.

  • type – The Type for this Variable.

  • value – A value to associate with this variable (a new container will be created).

  • strict – True -> assignments to self.value will not be casted or copied, so they must have the correct type or an exception will be raised.

  • container – The container to use for this variable. This should instead of the value parameter. Using both is an error.


A container to use for this SharedVariable when it is an implicit function parameter.

aesara.compile.sharedvalue.shared(value, name=None, strict=False, allow_downcast=None, **kwargs)[source]#

Create a SharedVariable initialized with a copy or reference of value.

This function iterates over constructor functions to find a suitable SharedVariable subclass. The suitable one is the first constructor that accept the given value. See the documentation of shared_constructor() for the definition of a constructor function.

This function is meant as a convenient default. If you want to use a specific constructor, consider calling it directly.

aesara.shared is a shortcut to this function.


By passing kwargs, you effectively limit the set of potential constructors to those that can accept those kwargs.

Some shared variable have borrow as a kwarg.

SharedVariables of TensorType have broadcastable as a kwarg. As shared variable shapes can change, all dimensions default to not being broadcastable, even if value has a shape of 1 along some dimension. This parameter allows one to create for example a row or column tensor.


Append ctor to the list of shared constructors (see shared()).

Each registered constructor ctor will be called like this:

ctor(value, name=name, strict=strict, **kwargs)

If it do not support given value, it must raise a TypeError.