Skip to content

Commit

Permalink
• transferfunctions.py:
Browse files Browse the repository at this point in the history 
  - rename bounds -> range
  - add DeterministicTransferFunction subclass of TransferFunction:
    - scale and offset Parameters used by all subclasses
    - add _range_setter() that adjusts range based on scale and/or offset
  • Loading branch information
@jdcpni
jdcpni committed Feb 1, 2025
1 parent c80991f commit 609a340
Show file tree
Hide file tree
Showing 6 changed files with 77 additions and 119 deletions.
2 changes: 1 addition & 1 deletion psyneulink/core/components/component.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1484,7 +1484,7 @@ def _get_compilation_params(self):
"control_signal", "competition",
"has_recurrent_input_port", "enable_learning",
"enable_output_type_conversion", "changes_shape",
"output_type", "bounds", "internal_only",
"output_type", "range", "internal_only",
"require_projection_in_composition", "default_input",
"shadow_inputs", "compute_reconfiguration_cost",
"reconfiguration_cost", "net_outcome", "outcome",
Expand Down
2 changes: 0 additions & 2 deletions psyneulink/core/components/functions/nonstateful/selectionfunctions.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -218,8 +218,6 @@ class OneHot(SelectionFunction):
function. Values specified for parameters in the dictionary override any assigned to those parameters in
arguments of the constructor.
bounds : None
owner : Component
`component <Component>` to which to assign the Function.
Expand Down
130 changes: 48 additions & 82 deletions psyneulink/core/components/functions/nonstateful/transferfunctions.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -40,7 +40,7 @@
Standard Attributes
~~~~~~~~~~~~~~~~~~~
All TransferFunctions have a `bounds <TransferFunction.bounds>` attribute that specifies the lower and upper limits
All TransferFunctions have a `range <TransferFunction.range>` attribute that specifies the lower and upper limits
of the function's result. For some subclasses, this may be modified by other parameters. In addition, all
TransferFunctions have a pair of modulable parameters as described below.
Expand Down Expand Up @@ -115,22 +115,22 @@
'Linear', 'Logistic', 'ReLU', 'SoftMax', 'Tanh', 'TransferFunction', 'TransferWithCosts'
]

def _bounds_getter_using_scale_and_offset(owning_component=None, context=None):
"""Reassign bounds based on scale and offset applied to function's output for default upper and lower bounds
def _range_getter_using_scale_and_offset(owning_component=None, context=None):
"""Reassign range based on scale and offset applied to function's default_range
"""
default_bounds = owning_component.default_bounds
default_range = owning_component.default_range
scale = owning_component.parameters.scale._get(context)
offset = owning_component.parameters.offset._get(context)

# Deal with lower bound = None:
lower_bound = -np.inf if default_bounds[0] is None else default_bounds[0]
lower_bound = -np.inf if default_range[0] is None else default_range[0]
output_for_fct_lower_bound = scale * lower_bound + offset

# Deal with upper bound = None:
upper_bound = np.inf if default_bounds[1] is None else default_bounds[1]
upper_bound = np.inf if default_range[1] is None else default_range[1]
output_for_fct_upper_bound = scale * upper_bound + offset

# Need to do this since scale could be negative, reversing upper and lower bounds:
# Need to do this since scale could be negative, reversing upper and lower range:
lower_bound = min(output_for_fct_lower_bound, output_for_fct_upper_bound)
upper_bound = max(output_for_fct_lower_bound, output_for_fct_upper_bound)

Expand All @@ -149,11 +149,14 @@ class TransferFunction(Function_Base):
Attributes
----------
bounds : tuple(lower bound: float, uppper bound: float)
range : tuple(lower bound: float, uppper bound: float)
read-only Parameter that indicates the lower and upper limits of the function's result. The two items of the
tuple indicate the lower and upper bounds, respectively, with `None` as the entry if there is no bound. Some
subclasses of TransferFunction may have other Parameters that influence the bounds, which are described under
the `bounds <TransferFunction.bounds>` attribute of the relevant subclass.
tuple indicate the lower and upper bounds of range, respectively, with `None` as the entry if there are no
bounds. Some subclasses of TransferFunction may have other Parameters that influence the range, which are
described under the `range <TransferFunction.range>` attribute of the relevant subclasses.
default_range : tuple(lower bound: float, uppper bound: float)
class attribute that indicates the upper and lower limits of the Function's result.
"""

componentType = TRANSFER_FUNCTION_TYPE
Expand All @@ -163,14 +166,14 @@ class Parameters(Function_Base.Parameters):
Attributes
----------
bounds
see `bounds <TransferFunction.bounds>`
range
see `range <TransferFunction.range>`
:default value: (None, None)
:type:
"""
bounds = Parameter((None,None), read_only=True)
range = Parameter((None,None), read_only=True)

def _gen_llvm_function_body(self, ctx, builder, params, state, arg_in, arg_out, *, tags:frozenset):
assert isinstance(arg_in.type.pointee, pnlvm.ir.ArrayType)
Expand Down Expand Up @@ -199,16 +202,19 @@ class DeterministicTransferFunction(TransferFunction):
In addition to the `standard attributes <TransferFunction_StandardAttributes>` of a TransferFunction,
all DeterministicTransferFunctions have a `scale <DeterministicTransferFunction.scale>` and `offset
<DeterministicTransferFunction.offset>` Parameter, that are used to determine the `bounds <TransferFunction.bounds>`
<DeterministicTransferFunction.offset>` Parameter, that are used to determine the `range <TransferFunction.range>`
Attributes
----------
bounds : tuple(lower bound: float, uppper bound: float)
read-only Parameter that indicates the lower and upper limits of the function's result, after the `scale
default_range : tuple(lower bound: float, uppper bound: float)
class attribute that indicates the upper and lower limits of the Function's result, before `scale
<DeterministicTransferFunction.scale>` or `offset <DeterministicTransferFunction.offset>` are applied.
range : tuple(lower bound: float, uppper bound: float)
read-only Parameter that indicates the lower and upper limits of the Function's result, after the `scale
<DeterministicTransferFunction.scale>` and `offset <DeterministicTransferFunction.offset>` Parameters
have been applied to the result of the function: :math:`result(lower, upper) * scale + offset`, where
result(lower, upper) is given by the default values of the `bounds <TransferFunction.bounds>` attribute.
have been applied to the Function's default_range: :math:`default_range(lower, upper) * scale + offset`.
scale : float
determines the value by which the result of the function is multiplied, before `offset
Expand All @@ -225,8 +231,8 @@ class Parameters(TransferFunction.Parameters):
Attributes
----------
bounds
see `bounds <TransferFunction.bounds>`
range
see `range <TransferFunction.range>`
:default value: None
:type:
Expand All @@ -243,25 +249,13 @@ class Parameters(TransferFunction.Parameters):
:default value: 0.0
:type: float
"""
bounds = Parameter((None, None),
getter=_bounds_getter_using_scale_and_offset,
range = Parameter((None, None),
getter=_range_getter_using_scale_and_offset,
read_only=True,
dependencies={'scale', 'offset'})
scale = Parameter(1.0, modulable=True)
offset = Parameter(0.0, modulable=True)

# # MODIFIED 1/29/25 OLD:
# @check_user_specified
# def __init__(self,
# offset: Optional[ValidParamSpecType] = None,
# scale: Optional[ValidParamSpecType] = None,
# **kwargs):
#
# super().__init__(offset=offset,
# scale=scale,
# **kwargs)
# # MODIFIED 1/29/25 END


# **********************************************************************************************************************
# Identity
Expand Down Expand Up @@ -328,7 +322,7 @@ class Identity(DeterministicTransferFunction): #
PREFERENCE_SET_NAME: 'IdentityClassPreferences',
REPORT_OUTPUT_PREF: PreferenceEntry(False, PreferenceLevel.INSTANCE),
}
default_bounds = (None, None)
default_range = (None, None)

@check_user_specified
@beartype
Expand Down Expand Up @@ -475,7 +469,7 @@ class Linear(DeterministicTransferFunction): #
value added to each element of `variable <Linear.variable>` after applying the `slope <Linear.slope>`
(if it is specified).
bounds : (None, None)
range : (None, None)
modified by `scale <Linear.scale> and/or `offset <Linear.offset>` if they are specified.
scale : float
Expand Down Expand Up @@ -506,7 +500,7 @@ class Linear(DeterministicTransferFunction): #

_model_spec_class_name_is_generic = True

default_bounds = (None, None)
default_range = (None, None)

class Parameters(DeterministicTransferFunction.Parameters):
"""
Expand Down Expand Up @@ -766,7 +760,7 @@ class Exponential(DeterministicTransferFunction): # ---------------------------
value added to `variable <Exponential.variable>` after multiplying by `rate <Exponential.rate>`
and before exponentiation; assigned as *ADDITIVE_PARAM* of the Exponential Function.
bounds : (0, None)
range : (0, None)
modified by `scale <Exponential.scale> and/or `offset <Exponential.offset>` if they are specified.
scale : float
Expand All @@ -790,7 +784,7 @@ class Exponential(DeterministicTransferFunction): # ---------------------------
"""

componentName = EXPONENTIAL_FUNCTION
default_bounds = (0, None)
default_range = (0, None)


class Parameters(DeterministicTransferFunction.Parameters):
Expand Down Expand Up @@ -1034,7 +1028,7 @@ class Logistic(DeterministicTransferFunction): # ------------------------------
value to add to each element of `variable <Logistic.variable>` before applying `gain <Logistic.gain>`;
this argument has an effect identical to bias, but with the opposite sign (see `note <Logistic_Note>` above).
bounds : (0, 1)
range : (0, 1)
modified by `scale <Gaussian.scale> and/or `offset <Gaussian.offset>` if they are specified.
scale : float
Expand All @@ -1060,7 +1054,7 @@ class Logistic(DeterministicTransferFunction): # ------------------------------
componentName = LOGISTIC_FUNCTION
# parameter_keywords.update({GAIN, BIAS})
_model_spec_class_name_is_generic = True
default_bounds = (0, 1)
default_range = (0, 1)


class Parameters(DeterministicTransferFunction.Parameters):
Expand Down Expand Up @@ -1346,7 +1340,7 @@ class Tanh(DeterministicTransferFunction): # ----------------------------------
value subtracted from each element of `variable <Tanh.variable>` before applying the `gain <Tanh.gain>`
(if it is specified). This attribute is identical to bias, with the opposite sign.
bounds : (None, None)
range : (None, None)
modified by `scale <TanH.scale> and/or `offset <TanH.offset>` if they are specified.
scale : float
Expand All @@ -1370,7 +1364,7 @@ class Tanh(DeterministicTransferFunction): # ----------------------------------

componentName = TANH_FUNCTION
# parameter_keywords.update({GAIN, BIAS, OFFSET})
default_bounds = (-1, 1)
default_range = (-1, 1)


class Parameters(DeterministicTransferFunction.Parameters):
Expand Down Expand Up @@ -1641,7 +1635,7 @@ class ReLU(DeterministicTransferFunction): # ----------------------------------
leak : float : default 0.0
scaling factor between 0 and 1 when (variable - bias) is less than or equal to 0.
bounds : (None, None)
range : (None, None)
modified by `scale <Gaussian.scale> and/or `offset <ReLU.offset>` if they are specified.
scale : float : default 1.0
Expand All @@ -1665,7 +1659,7 @@ class ReLU(DeterministicTransferFunction): # ----------------------------------

componentName = RELU_FUNCTION
# parameter_keywords.update({GAIN, BIAS, LEAK})
default_bounds = (None, None)
default_range = (None, None)


class Parameters(DeterministicTransferFunction.Parameters):
Expand Down Expand Up @@ -1917,7 +1911,7 @@ class Gaussian(DeterministicTransferFunction): # ------------------------------
bias : float : default 0.0
value added to each element of `variable <Gaussian.variable>` before applying the Gaussian transform.
bounds : (None, None)
range : (None, None)
modified by `scale <Gaussian.scale> and/or `offset <Gaussian.offset>` if they are specified.
scale : float
Expand All @@ -1942,7 +1936,7 @@ class Gaussian(DeterministicTransferFunction): # ------------------------------

componentName = GAUSSIAN_FUNCTION
# parameter_keywords.update({STANDARD_DEVIATION, BIAS, SCALE, OFFSET})
default_bounds = (None, None)
default_range = (None, None)

class Parameters(TransferFunction.Parameters):
"""
Expand Down Expand Up @@ -2238,7 +2232,7 @@ class Parameters(TransferFunction.Parameters):
offset = Parameter(0.0, modulable=True)
random_state = Parameter(None, loggable=False, getter=_random_state_getter, dependencies='seed')
seed = Parameter(DEFAULT_SEED(), modulable=True, fallback_default=True, setter=_seed_setter)
bounds = (None, None)
range = (None, None)

@check_user_specified
@beartype
Expand Down Expand Up @@ -2462,7 +2456,7 @@ class Parameters(TransferFunction.Parameters):
p = Parameter(0.5, modulable=True, aliases=[MULTIPLICATIVE_PARAM])
random_state = Parameter(None, loggable=False, getter=_random_state_getter, dependencies='seed')
seed = Parameter(DEFAULT_SEED(), modulable=True, fallback_default=True, setter=_seed_setter)
bounds = (None, None)
range = (None, None)

@check_user_specified
@beartype
Expand Down Expand Up @@ -2959,7 +2953,7 @@ class SoftMax(TransferFunction):
for 2d variables, determines whether the SoftMax function is applied to the entire variable (per_item =
False), or applied to each item in the variable separately (per_item = True).
bounds : None if `output <SoftMax.output>` in {ARG_MAX, MAX_VAL}, else (0, 1) : default (0, 1)
range : None if `output <SoftMax.output>` in {ARG_MAX, MAX_VAL}, else (0, 1) : default (0, 1)
owner : Component
`component <Component>` to which the Function has been assigned.
Expand Down Expand Up @@ -3006,8 +3000,8 @@ class Parameters(TransferFunction.Parameters):
:default value: 0.1
:type: ``float``
bounds
see `bounds <SoftMax.bounds>`
range
see `range <SoftMax.range>`
:default value: (0, 1)
:type: <class 'tuple'>
Expand Down Expand Up @@ -3042,7 +3036,7 @@ class Parameters(TransferFunction.Parameters):
adapt_scale = Parameter(1.0, modulable=True)
adapt_base = Parameter(1.0, modulable=True)
adapt_entropy_weighting = Parameter(0.95, modulable=True)
bounds = (0, 1)
range = (0, 1)
output = ALL
per_item = Parameter(True, pnl_internal=True)
one_hot_function = Parameter(None, stateful=False, loggable=False)
Expand Down Expand Up @@ -4460,13 +4454,6 @@ def _function(self,
# Compute current intensity
intensity = self.parameters.transfer_fct._get(context)(variable, context=context)

# # MODIFIED 1/29/25 OLD:
# # Apply scale and offset to current intensity (which is value returned by function)
# scale = self.parameters.scale._get(context)
# offset = self.parameters.offset._get(context)
# intensity = scale * intensity + offset
# MODIFIED 1/29/25 END

# THEN, DEAL WITH COSTS
# Note: only compute costs that are enabled; others are left as None, or with their value when last enabled.

Expand Down Expand Up @@ -4525,16 +4512,7 @@ def _is_identity(self, context=None, defaults=False):
else:
enabled_cost_functions = self.parameters.enabled_cost_functions.get(context)

# MODIFIED 1/29/25 OLD:
return transfer_fct._is_identity(context, defaults=defaults) and enabled_cost_functions == CostFunctions.NONE
# MODIFIED 1/29/25 NEW:
# return (
# transfer_fct._is_identity(context, defaults=defaults)
# and self.parameters.scale.get(context) == 1.0
# and self.parameters.offset.get(context) == 0.0
# and enabled_cost_functions == CostFunctions.NONE
# )
# MODIFIED 1/29/25 END

@beartype
def assign_costs(self, cost_functions: Union[CostFunctions, list], execution_context=None):
Expand Down Expand Up @@ -4660,18 +4638,6 @@ def _gen_llvm_function_body(self, ctx, builder, params, state, arg_in, arg_out,
trans_out = arg_out
builder.call(trans_f, [trans_p, trans_s, trans_in, trans_out])

# # MODIFIED 1/29/25 NEW:
# # apply scale and offset
# scale_ptr = ctx.get_param_or_state_ptr(builder, self, SCALE, param_struct_ptr=params)
# offset_ptr = ctx.get_param_or_state_ptr(builder, self, OFFSET, param_struct_ptr=params)
# scale = pnlvm.helpers.load_extract_scalar_array_one(builder, scale_ptr)
# offset = pnlvm.helpers.load_extract_scalar_array_one(builder, offset_ptr)
# # trans_out = pnlvm.helpers.load_extract_scalar_array_one(builder, trans_out)
# # trans_out = builder.fmul(trans_out, scale)
# # trans_out = builder.fadd(trans_out, offset)
# # builder.store(builder.load(trans_out), trans_out)
# # MODIFIED 1/29/25 END

intensity_ptr = ctx.get_state_space(builder, self, state, self.parameters.intensity)

costs = [(self.parameters.intensity_cost_fct, CostFunctions.INTENSITY, self.parameters.intensity_cost),
Expand Down
3 changes: 0 additions & 3 deletions psyneulink/core/components/functions/nonstateful/transformfunctions.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1735,8 +1735,6 @@ class MatrixTransform(TransformFunction): # -----------------------------------
be used if `variable <MatrixTransform.variable>` is a scalar (i.e., has only one element), and **operation**
is set to *L0* (since it is not needed, and can produce a divide by zero error).
bounds : None
params : Dict[param keyword: param value] : default None
a `parameter dictionary <ParameterPort_Specification>` that specifies the parameters for the
function. Values specified for parameters in the dictionary override any assigned to those parameters in
Expand Down Expand Up @@ -1823,7 +1821,6 @@ class Parameters(TransformFunction.Parameters):
matrix = Parameter(None, modulable=True, mdf_name='B')
operation = Parameter(DOT_PRODUCT, stateful=False)
normalize = Parameter(False)
bounds = None

@check_user_specified
@beartype
Expand Down
Loading

0 comments on commit 609a340

Please sign in to comment.