提交 fa469ebd authored 作者: Razvan Pascanu's avatar Razvan Pascanu

changing the naming conventions in the scan function

上级 c066deb4
...@@ -58,7 +58,7 @@ scan op implements system of equations of the following form: ...@@ -58,7 +58,7 @@ scan op implements system of equations of the following form:
The equations describe a system evolving in time, where :math:`t` represents the The equations describe a system evolving in time, where :math:`t` represents the
current step. The system is described by inputs, states, outputs and current step. The system is described by inputs, states, outputs and
weights. parameteres.
The inputs, denoted by :math:`\mathbf{u}` are time-varying quantities, The inputs, denoted by :math:`\mathbf{u}` are time-varying quantities,
hence indexed by :math:`t`. They however only influence the system, but are hence indexed by :math:`t`. They however only influence the system, but are
...@@ -66,7 +66,7 @@ not influenced by the system. ...@@ -66,7 +66,7 @@ not influenced by the system.
The states :math:`\mathbf{x}` are time-varying quantities, whose value at The states :math:`\mathbf{x}` are time-varying quantities, whose value at
time :math:`t` depends on its (or other state) previous values as well as time :math:`t` depends on its (or other state) previous values as well as
the inputs and weights. Note that the first few values of the states are the inputs and parameters. Note that the first few values of the states are
always provided, otherwise we could not imploy the recurrent equation to always provided, otherwise we could not imploy the recurrent equation to
generate these sequence of values without a starting point. generate these sequence of values without a starting point.
...@@ -74,7 +74,7 @@ The outputs, :math:`\mathbf{y}` are outputs of the system, i.e. values that ...@@ -74,7 +74,7 @@ The outputs, :math:`\mathbf{y}` are outputs of the system, i.e. values that
depend on the previous values of the states and inputs. The difference depend on the previous values of the states and inputs. The difference
between outputs and states is that outputs do not feed back into the system. between outputs and states is that outputs do not feed back into the system.
The weights :math:`\mathbf{w}` are fixed quantities that are re-used at The parameters :math:`\mathbf{w}` are fixed quantities that are re-used at
every time step of the evolution of the system. every time step of the evolution of the system.
Each of the equations above are implemented by the **inner function** of scan. You Each of the equations above are implemented by the **inner function** of scan. You
...@@ -88,24 +88,19 @@ computational graph that is afterwards compiled into the **inner function**. ...@@ -88,24 +88,19 @@ computational graph that is afterwards compiled into the **inner function**.
Naming conventions Naming conventions
------------------ ------------------
* ``x`` will stand for a state :math:`\mathbf{x}`, while ``xs`` will represent * ``input_state`` will stand for a state :math:`\mathbf{x}`, when it is
the list of all states provided as an input to the recurrent formula (the inner function) that
* ``y`` will stand for an output :math:`\mathbf{y}`, while ``ys`` will will generate the new value of the state
represent the list of all outputs * ``output_state`` will stand for a state :math:`\math{x}` when it refers
* ``xy`` will stand for either a state or an output, while ``xys`` will be to the result of the recurrent formula (the output of the inner function)
the list of all states and outputs * ``output`` will stand for an output :math:`\mathbf{y}`
* ``u`` will be an input, wile ``us`` will be the list of all inputs * ``input`` will be an input :math:`\mathbf{u}`
* ``w`` will stand for a weight tensor, while ``ws`` for the list of all * ``parameter`` will stand for a parameter tensor :math:`\mathbf{w}` that stays
weight tensors constant at each step of the inner function
* ``z`` will stand for states that are not numeric in nature. More * ``shared_parameter`` will stand for states that are not numeric in nature. More
specifically *random states*. ``zs`` is the list of all such states. specifically *random states*.
* ``t`` is the time index (the current step in the evolution of the system). * ``t`` is the time index (the current step in the evolution of the system).
* ``T`` is the total number of steps in the evolution of the system. * ``T`` is the total number of steps in the evolution of the system.
* ``N`` will be the number of states, ``n`` will be the index that loops
over the number of states
* ``M`` will be the number of inputs and ``m`` the dedicated index for
inputs
* ``Q`` the number of weights
* the suffix ``_slices`` added to either ``x`` or ``u`` will mean the list of * the suffix ``_slices`` added to either ``x`` or ``u`` will mean the list of
variables representing slices of states or inputs. These are the arguments variables representing slices of states or inputs. These are the arguments
given to the constructive function of scan (see above). given to the constructive function of scan (see above).
...@@ -115,9 +110,6 @@ Naming conventions ...@@ -115,9 +110,6 @@ Naming conventions
* the suffix ``_outer`` added to ``x``, ``y``, ``xy``, ``u``, ``w`` or ``z`` * the suffix ``_outer`` added to ``x``, ``y``, ``xy``, ``u``, ``w`` or ``z``
will mean the variables representing the state/output/input/weights in the will mean the variables representing the state/output/input/weights in the
main computational graph (the one containing the scan op). main computational graph (the one containing the scan op).
* the suffx ``_results`` added to ``x``, ``y`` or ``xy`` are the list of
expressions representing the next value of the state/output (i.e. the
results after executing the inner function of scan)
Files Files
----- -----
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论