chainer.FunctionNode¶

class
chainer.
FunctionNode
[source]¶ Function node of the computational graph.
FunctionNode is a class representing a node in a computational graph. The node corresponds to an application of a differentiable function to input variables.
When a differentiable function is applied to
Variable
objects, it creates an instance of FunctionNode implementation and calls itsapply()
method. Theapply()
method basically does the following three things. Adding an edge from the function node to the variable node corresponding
to each input. The node of each input is extracted by
Variable.node
.  Computing the output arrays of the function.
 Creating a
Variable
object for each output array and adding an edge from the node of the variable to the function node.
The output variables are then returned.
Example
Let
x
be an instance ofVariable
andf
be an instance ofFunctionNode
taking only one argument. Then the following code>>> import numpy, chainer, chainer.functions as F >>> x = chainer.Variable(numpy.zeros(10)) >>> f = F.Identity() >>> y = f.apply((x,))[0]
computes a new variable
y
and creates backward references. The backward references are actually set as per the following diagram:x.node < f < y.node
If an application of another function
g
occurs as>>> g = F.Identity() >>> z = g.apply((x,))[0]
then the graph grows with a branch:
 f < y.node x.node <+  g < z.node
Note that the branching is correctly managed on backward computation, i.e. the gradients from
f
andg
are accumulated to the gradient ofx
.Every functionnode implementation should provide
forward()
andbackward()
. Instead of overridingforward()
, one can also implementforward_cpu()
andforward_gpu()
when the implementations for CPU and GPU arrays are totally different.Note that the input and output variables are inaccessible from
backward()
by default. If it needs accesses to these variables, theforward()
method (or its CPU/GPU variants) has to callretain_inputs()
andretain_outputs()
appropriately. The retained input/output variables can be accessed frombackward()
by callingget_retained_inputs()
andget_retained_outputs()
.Note
There are two types of differentiable functions in Chainer (since v3). The first type is of a function using a subclass of
Function
, which is called oldstyle differentiable function. The second type is of a function using a subclass ofFunctionNode
, which is called newstyle differentiable function. There are several advantages on using the newstyle differentiable function. The newstyle differentiable function supports differentiable backpropagation. The backpropagated gradients computed through the newstyle differentiable functions themselves support further backpropagations so that the automatic higherorder differentiation is available.
 The backpropagation of the newstyle differentiable function can be more computationally efficient because the interface allows an implementation to omit the computation of unneeded input gradients.
Note that the newstyle differentiable function is the standard way of defining a function node of the computational graph in Chainer; old style differentiable functions are implemented as wrappers of the new style differentiable functions.
Variables:  inputs – A tuple of the input
VariableNode
objects.  outputs – A tuple of weak references to the output
VariableNode
objects.  rank (int) – An ordinal following the topological order of the computational graph.
 stack – Stack trace retrieved at the forward computation. The stack trace is available only in the debug mode.
New in version 3.0.0.
Methods

add_hook
(hook, name=None)[source]¶ Registers a function hook.
Parameters:  hook (FunctionHook) – Function hook to be registered.
 name (str) – Name of the function hook. The name must be unique
among function hooks registered to this function. If
None
, the default name of the function hook is used.

apply
(inputs)[source]¶ Computes output variables and grows the computational graph.
Basic behavior is expressed in the documentation of
FunctionNode
.Note
If the
data
attribute of input variables exist on a GPU device, that device is made current before callingforward()
, so implementors do not need to take care of device selection in most cases.Parameters: inputs – Tuple of input variables. Each element can be either Variable
,numpy.ndarray
, orcupy.ndarray
. If the element is an ndarray, it is automatically wrapped withVariable
.Returns: A tuple of output Variable
objects.

backward
(target_input_indexes, grad_outputs)[source]¶ Computes gradients w.r.t. specified inputs given output gradients.
This method is used to compute one step of the backpropagation corresponding to the forward computation of this function node. Given the gradients w.r.t. output variables, this method computes the gradients w.r.t. specified input variables. Note that this method does not need to compute any input gradients not specified by
target_input_indices
.Unlike
Function.backward()
, gradients are given asVariable
objects and this method itself has to return input gradients asVariable
objects. It enables the function node to return the input gradients with the full computational history, in which case it supports differentiable backpropagation or higherorder differentiation.The default implementation returns
None
s, which means the function is not differentiable.Parameters:  target_input_indexes (tuple of int) – Indices of the input variables w.r.t. which the gradients are required. It is guaranteed that this tuple contains at least one element.
 grad_outputs (tuple of
Variable
s) – Gradients w.r.t. the output variables. If the gradient w.r.t. an output variable is not given, the corresponding element isNone
.
Returns: Tuple of variables that represent the gradients w.r.t. specified input variables. The length of the tuple can be same as either
len(target_input_indexes)
or the number of inputs. In the latter case, the elements not specified bytarget_input_indexes
will be discarded.See also
backward_accumulate()
provides an alternative interface that allows you to implement the backward computation fused with the gradient accumulation.

backward_accumulate
(target_input_indexes, grad_outputs, grad_inputs)[source]¶ Computes gradients w.r.t. specified inputs and accumulates them.
This method provides a way to fuse the backward computation and the gradient accumulations in the case that the multiple functions are applied to the same variable.
Users have to override either of this method or
backward()
. It is often simpler to implementbackward()
and is recommended if you do not need to provide efficient gradient accumulation.Parameters:  target_input_indexes (tuple of int) – Indices of the input variables w.r.t. which the gradients are required. It is guaranteed that this tuple contains at least one element.
 grad_outputs (tuple of Variable) – Gradients w.r.t. the output
variables. If the gradient w.r.t. an output variable is not
given, the corresponding element is
None
.  grad_inputs (tuple of Variable) – Gradients w.r.t. the input
variables specified by
target_input_indexes
. These values are computed by other computation paths. If there is no gradient value existing for the variable, the corresponding element isNone
. See also the note below.
Returns: Tuple of variables that represent the gradients w.r.t. specified input variables. Unlike
backward()
, the length of the tuple must be same as that oftarget_input_indices
.Note
When the same variable is passed to the multiple input arguments of a function, only the first position of
grad_inputs
corresponding to these input arguments may contain the gradient variable corresponding to that input variable, and other entries are set toNone
. This is an implementationdetail convention to avoid the complication of correctly accumulating gradients in such a case. This behavior might be changed in a future version.

check_type_forward
(in_types)[source]¶ Checks types of input data before forward propagation.
This method is called before
forward()
and validates the types of input variables using the type checking utilities.Parameters: in_types (TypeInfoTuple) – The type information of input variables for forward()
.

delete_hook
(name)[source]¶ Unregisters the function hook.
Parameters: name (str) – The name of the function hook to be unregistered.

forward
(inputs)[source]¶ Computes the output arrays from the input arrays.
It delegates the procedure to
forward_cpu()
orforward_gpu()
by default. Which of them this method selects is determined by the type of input arrays. Implementations ofFunctionNode
must implement either CPU/GPU methods or this method.Parameters: inputs – Tuple of input array(s). Returns: Tuple of output array(s). Warning
Implementations of
FunctionNode
must take care that the return value must be a tuple even if it returns only one array.

forward_cpu
(inputs)[source]¶ Computes the output arrays from the input NumPy arrays.
Parameters: inputs – Tuple of input numpy.ndarray
objects.Returns: Tuple of output arrays. Each element can be NumPy or CuPy arrays. Warning
Implementation of
FunctionNode
must take care that the return value must be a tuple even if it returns only one array.

forward_gpu
(inputs)[source]¶ Computes the output arrays from the input CuPy arrays.
Parameters: inputs – Tuple of input cupy.ndarray
objects.Returns: Tuple of output arrays. Each element can be NumPy or CuPy arrays. Warning
Implementation of
FunctionNode
must take care that the return value must be a tuple even if it returns only one array.

get_retained_inputs
()[source]¶ Returns a tuple of retained input variables.
This method is used to retrieve the input variables retained in
forward()
.Returns: A tuple of retained input variables.

get_retained_outputs
()[source]¶ Returns a tuple of retained output variables.
This method is used to retrieve the output variables retained in
forward()
.Returns: A tuple of retained output variables. Note
This method does a tricky thing to support the case of an output node garbagecollected before this method is called; in this case, this method creates a fresh variable node that acts as an output node of the function node.

retain_inputs
(indexes)[source]¶ Lets specified input variable nodes keep data arrays.
By calling this method from
forward()
, the function node can specify which inputs are required for backprop. The input variables with retained arrays can then be obtained by callingget_retained_inputs()
from insidebackward()
.Unlike
Function
, the function node DOES NOT keep input arrays by default. If you want to keep some or all input arrays, do not forget to call this method.Note that this method must not be called from the outside of
forward()
.Parameters: indexes (iterable of int) – Indexes of input variables that the function will require for backprop.

retain_outputs
(indexes)[source]¶ Lets specified output variable nodes keep data arrays.
By calling this method from
forward()
, the function node can specify which outputs are required for backprop. If this method is not called, no output variables will be marked to keep their data array at the point of returning fromapply()
. The output variables with retained arrays can then be obtained by callingget_retained_outputs()
from insidebackward()
.Note
It is recommended to use this method if the function requires some or all output arrays in backprop. The function can also use output arrays just by keeping references to them directly, although it might affect the performance of later function applications on the output variables.
Note that this method must not be called from the outside of
forward()
.Parameters: indexes (iterable of int) – Indexes of output variables that the function will require for backprop.
Attributes

inputs
= None¶

label
¶ Short text that represents the function.
The default implementation returns its type name. Each function should override it to give more information.

local_function_hooks
¶ Ordered dictionary of registered function hooks.
Contrary to
chainer.thread_local.function_hooks
, which registers its elements to all functions, Function hooks in this property is specific to this function.

output_data
¶ A tuple of the retained output arrays.
This property is mainly used by
Function
. Users basically do not have to use this property; useget_retained_outputs()
instead.

outputs
= None¶

rank
= 0¶

stack
= None¶
 Adding an edge from the function node to the variable node corresponding
to each input. The node of each input is extracted by