class Parameters(Function.Parameters):
"""
Attributes
----------
variable
see `variable <Function_Base.variable>`
:default value: numpy.array([0])
:type: ``numpy.ndarray``
:read only: True
enable_output_type_conversion
see `enable_output_type_conversion <Function_Base.enable_output_type_conversion>`
:default value: False
:type: ``bool``
output_type
see `output_type <Function_Base.output_type>`
:default value: FunctionOutputType.DEFAULT
:type: `FunctionOutputType`
"""
variable = Parameter(np.array([0]), read_only=True, pnl_internal=True, constructor_argument='default_variable')
output_type = Parameter(
FunctionOutputType.DEFAULT,
stateful=False,
loggable=False,
pnl_internal=True,
valid_types=FunctionOutputType
)
enable_output_type_conversion = Parameter(False, stateful=False, loggable=False, pnl_internal=True)
class Parameters(ModulatoryProjection_Base.Parameters):
"""
Attributes
----------
control_signal
see `control_signal <ControlProjection.control_signal>`
:default value: None
:type:
:read only: True
function
see `function <ControlProjection.function>`
:default value: `Linear`
:type: `Function`
"""
function = Parameter(Linear, stateful=False, loggable=False)
control_signal = Parameter(None, read_only=True, getter=_control_signal_getter, setter=_control_signal_setter, pnl_internal=True)
control_signal_params = Parameter(
None,
stateful=False,
loggable=False,
read_only=True,
user=False,
pnl_internal=True
)
class Parameters(ObjectiveFunction.Parameters):
"""
Attributes
----------
matrix
see `matrix <Stability.matrix>`
:default value: `HOLLOW_MATRIX`
:type: ``str``
metric
see `metric <Stability.metric>`
:default value: `ENERGY`
:type: ``str``
metric_fct
see `metric_fct <Stability.metric_fct>`
:default value: None
:type:
transfer_fct
see `transfer_fct <Stability.transfer_fct>`
:default value: None
:type:
"""
matrix = HOLLOW_MATRIX
metric = Parameter(ENERGY, stateful=False)
metric_fct = Parameter(None, stateful=False, loggable=False)
transfer_fct = Parameter(None, stateful=False, loggable=False)
normalize = Parameter(False, stateful=False)
class Parameters(ControlMechanism.Parameters):
"""
Attributes
----------
value
see `value <GatingMechanism.value>`
:default value: numpy.array([0.5])
:type: ``numpy.ndarray``
gating_allocation
see `gating_allocation <GatingMechanism.gating_allocation>`
:default value: numpy.array([0.5])
:type: ``numpy.ndarray``
:read only: True
"""
# This must be a list, as there may be more than one (e.g., one per control_signal)
value = Parameter(np.array([defaultGatingAllocation]),
aliases='control_allocation',
pnl_internal=True)
gating_allocation = Parameter(np.array([defaultGatingAllocation]),
getter=_gating_allocation_getter,
setter=_gating_allocation_setter,
read_only=True,
pnl_internal=True)
class Parameters(IntegratorMechanism.Parameters):
"""
Attributes
----------
filter_function
see `filter_function <PredictionMechanism.filter_function>`
:default value: None
:type:
rate
see `rate <PredictionMechanism.rate>`
:default value: 1.0
:type: float
window_size
see `window_size <PredictionMechanism.window_size>`
:default value: 1
:type: int
"""
window_size = Parameter(1, stateful=False, loggable=False)
filter_function = Parameter(None, stateful=False, loggable=False)
input_type = None
rate = Parameter(1.0, modulable=True)
class Parameters(DistributionFunction.Parameters):
"""
Attributes
----------
variable
see `variable <UniformToNormalDist.variable>`
:default value: numpy.array([0])
:type: numpy.ndarray
:read only: True
mean
see `mean <UniformToNormalDist.mean>`
:default value: 0.0
:type: float
standard_deviation
see `standard_deviation <UniformToNormalDist.standard_deviation>`
:default value: 1.0
:type: float
"""
variable = Parameter(np.array([0]),
read_only=True,
pnl_internal=True,
constructor_argument='default_variable')
mean = Parameter(0.0, modulable=True, aliases=[ADDITIVE_PARAM])
standard_deviation = Parameter(1.0,
modulable=True,
aliases=[MULTIPLICATIVE_PARAM])
class Parameters(SelectionFunction.Parameters):
"""
Attributes
----------
mode
see `mode <OneHot.mode>`
:default value: `MAX_VAL`
:type: ``str``
random_state
see `random_state <OneHot.random_state>`
:default value: None
:type: ``numpy.random.RandomState``
"""
mode = Parameter(MAX_VAL, stateful=False)
random_state = Parameter(None, stateful=True, loggable=False)
def _validate_mode(self, mode):
options = {
MAX_VAL, MAX_ABS_VAL, MAX_INDICATOR, MAX_ABS_INDICATOR,
MIN_VAL, MIN_ABS_VAL, MIN_INDICATOR, MIN_ABS_INDICATOR, PROB,
PROB_INDICATOR
}
if mode in options:
# returns None indicating no error message (this is a valid assignment)
return None
else:
# returns error message
return 'not one of {0}'.format(options)
class Parameters(ControlMechanism.Parameters):
"""
Attributes
----------
base_level_gain
see `base_level_gain <LCControlMechanism.base_level_gain>`
:default value: 0.5
:type: ``float``
function
see `function <LCControlMechanism_Function>`
:default value: `FitzHughNagumoIntegrator`
:type: `Function`
modulated_mechanisms
see `modulated_mechanisms <LCControlMechanism_Modulated_Mechanisms>`
:default value: None
:type:
scaling_factor_gain
see `scaling_factor_gain <LCControlMechanism.scaling_factor_gain>`
:default value: 3.0
:type: ``float``
"""
function = Parameter(FitzHughNagumoIntegrator, stateful=False, loggable=False)
base_level_gain = Parameter(0.5, modulable=True)
scaling_factor_gain = Parameter(3.0, modulable=True)
modulated_mechanisms = Parameter(None, stateful=False, loggable=False)
class Parameters(ProcessingMechanism_Base.Parameters):
"""
Attributes
----------
variable
see `variable <Mechanism_Base.variable>`
:default value: [[0]]
:type: list
"""
variable = Parameter([[0]], pnl_internal=True, constructor_argument='default_variable')
function = Parameter(ContentAddressableMemory, stateful=False, loggable=False)
content_size = 1
assoc_size = 0
input_ports = Parameter(
_generate_content_input_port_spec(content_size),
stateful=False,
loggable=False,
read_only=True,
structural=True,
parse_spec=True,
)
output_ports = Parameter(
[{NAME: CONTENT_OUTPUT, VARIABLE: (OWNER_VALUE, 0)}],
stateful=False,
loggable=False,
read_only=True,
structural=True,
)
class Parameters(LearningMechanism.Parameters):
"""
Attributes
----------
modulation
see `modulation <AutoAssociativeLearningMechanism.modulation>`
:default value: ADDITIVE
:type: str
"""
function = Parameter(Hebbian, stateful=False, loggable=False)
modulation = ADDITIVE
input_ports = Parameter(
[ACTIVATION_INPUT],
stateful=False,
loggable=False,
read_only=True,
structural=True,
parse_spec=True,
)
output_ports = Parameter(
[{
NAME:
LEARNING_SIGNAL, # NOTE: This is the default, but is overridden by any LearningSignal arg
VARIABLE: (OWNER_VALUE, 0)
}],
stateful=False,
loggable=False,
read_only=True,
structural=True,
)
class Parameters(Function_Base.Parameters):
"""
Attributes
----------
initializer
see `initializer <StatefulFunction.initializer>`
:default value: numpy.array([0])
:type: numpy.ndarray
noise
see `noise <StatefulFunction.noise>`
:default value: 0.0
:type: float
previous_value
see `previous_value <StatefulFunction.previous_value>`
:default value: numpy.array([0])
:type: numpy.ndarray
rate
see `rate <StatefulFunction.rate>`
:default value: 1.0
:type: float
"""
noise = Parameter(0.0, modulable=True)
rate = Parameter(1.0, modulable=True)
previous_value = Parameter(np.array([0]), pnl_internal=True)
initializer = Parameter(np.array([0]), pnl_internal=True)
class Parameters(ComparatorMechanism.Parameters):
"""
Attributes
----------
variable
see `variable <Mechanism_Base.variable>`
:default value: None
:type:
:read only: True
function
see `function <PredictionErrorMechanism.function>`
:default value: `PredictionErrorDeltaFunction`
:type: `Function`
learning_rate
see `learning_rate <PredictionErrorMechanism.learning_rate>`
:default value: 0.3
:type: float
"""
variable = Parameter(None, read_only=True, pnl_internal=True, constructor_argument='default_variable')
learning_rate = Parameter(0.3, modulable=True)
function = PredictionErrorDeltaFunction
sample = None
target = None
class Parameters(ModulatorySignal.Parameters):
"""
Attributes
----------
variable
see `variable <ControlSignal.variable>`
:default value: numpy.array([1.])
:type: numpy.ndarray
value
see `value <GatingSignal.value>`
:default value: numpy.array([0])
:type: numpy.ndarray
:read only: True
allocation_samples
see `allocation_samples <ControlSignal.allocation_samples>`
"""
variable = Parameter(np.array(defaultGatingAllocation),
aliases='allocation',
getter=_output_state_variable_getter)
value = Parameter(np.array(defaultGatingAllocation), read_only=True, aliases=['intensity'])
allocation_samples = Parameter(np.arange(0.1, 1.01, 0.3), modulable=True)
class Parameters(ModulatoryProjection_Base.Parameters):
"""
Attributes
----------
function
see `function <Projection_Base.function>`
:default value: `Linear`
:type: `Function`
gating_signal
see `gating_signal <GatingProjection.gating_signal>`
:default value: None
:type:
:read only: True
"""
function = Parameter(Linear(params={FUNCTION_OUTPUT_TYPE: FunctionOutputType.RAW_NUMBER}), stateful=False, loggable=False)
gating_signal = Parameter(None, read_only=True, getter=_gating_signal_getter, setter=_gating_signal_setter, pnl_internal=True)
gating_signal_params = Parameter(
None,
stateful=False,
loggable=False,
read_only=True,
user=False,
pnl_internal=True
)
class Parameters(RecurrentTransferMechanism.Parameters):
"""
Attributes
----------
average_based
see `average_based <KWTAMechanism.average_based>`
:default value: False
:type: ``bool``
function
see `function <KWTAMechanism.function>`
:default value: `Logistic`
:type: `Function`
inhibition_only
see `inhibition_only <KWTAMechanism.inhibition_only>`
:default value: True
:type: ``bool``
k_value
see `k_value <KWTAMechanism.k_value>`
:default value: 0.5
:type: ``float``
ratio
see `ratio <KWTAMechanism.ratio>`
:default value: 0.5
:type: ``float``
threshold
see `threshold <KWTAMechanism.threshold>`
:default value: 0.0
:type: ``float``
"""
function = Parameter(Logistic,
stateful=False,
loggable=False,
dependencies='integrator_function')
k_value = Parameter(0.5, modulable=True)
threshold = Parameter(0.0, modulable=True)
ratio = Parameter(0.5, modulable=True)
output_ports = Parameter(
[RESULT],
stateful=False,
loggable=False,
read_only=True,
structural=True,
)
average_based = False
inhibition_only = True
class Parameters(ProcessingMechanism_Base.Parameters):
"""
Attributes
----------
hidden_layers
see `hidden_layers <LeabraMechanism.hidden_layers>`
:default value: 0
:type: ``int``
hidden_sizes
see `hidden_sizes <LeabraMechanism.hidden_sizes>`
:default value: None
:type:
input_size
see `input_size <LeabraMechanism.input_size>`
:default value: 1
:type: ``int``
network
see `network <LeabraMechanism.network>`
:default value: None
:type:
output_size
see `output_size <LeabraMechanism.output_size>`
:default value: 1
:type: ``int``
quarter_size
see `quarter_size <LeabraMechanism.quarter_size>`
:default value: 50
:type: ``int``
training_flag
see `training_flag <LeabraMechanism.training_flag>`
:default value: None
:type:
"""
input_size = 1
output_size = 1
hidden_layers = 0
hidden_sizes = None
quarter_size = 50
function = Parameter(LeabraFunction, stateful=False, loggable=False)
network = FunctionParameter(None)
training_flag = Parameter(False,
setter=_training_flag_setter,
dependencies='network')
class Parameters(ProcessingMechanism_Base.Parameters):
"""
Attributes
----------
hidden_layers
see `hidden_layers <LeabraMechanism.hidden_layers>`
:default value: 0
:type: int
hidden_sizes
see `hidden_sizes <LeabraMechanism.hidden_sizes>`
:default value: None
:type:
input_size
see `input_size <LeabraMechanism.input_size>`
:default value: 1
:type: int
network
see `network <LeabraMechanism.network>`
:default value: None
:type:
output_size
see `output_size <LeabraMechanism.output_size>`
:default value: 1
:type: int
quarter_size
see `quarter_size <LeabraMechanism.quarter_size>`
:default value: 50
:type: int
training_flag
see `training_flag <LeabraMechanism.training_flag>`
:default value: None
:type:
"""
input_size = 1
output_size = 1
hidden_layers = 0
hidden_sizes = None
quarter_size = 50
network = Parameter(None,
getter=_network_getter,
setter=_network_setter)
training_flag = Parameter(None, setter=_training_flag_setter)
class Parameters(MappingProjection.Parameters):
"""
Attributes
----------
variable
see `variable <AutoAssociativeProjection.variable>`
:default value: numpy.array([[0]])
:type: numpy.ndarray
:read only: True
auto
see `auto <AutoAssociativeProjection.auto>`
:default value: 1
:type: int
function
see `function <AutoAssociativeProjection.function>`
:default value: `LinearMatrix`
:type: `Function`
hetero
see `hetero <AutoAssociativeProjection.hetero>`
:default value: 0
:type: int
matrix
see `matrix <AutoAssociativeProjection.matrix>`
:default value: `AUTO_ASSIGN_MATRIX`
:type: str
"""
variable = Parameter(np.array([[0]]),
read_only=True,
pnl_internal=True,
constructor_argument='default_variable')
# function is always LinearMatrix that requires 1D input
function = Parameter(LinearMatrix, stateful=False, loggable=False)
auto = Parameter(1,
getter=_auto_getter,
setter=_auto_setter,
modulable=True)
hetero = Parameter(0,
getter=_hetero_getter,
setter=_hetero_setter,
modulable=True)
matrix = Parameter(DEFAULT_MATRIX,
function_parameter=True,
getter=_matrix_getter,
setter=_matrix_setter,
modulable=True)
def _instantiate_attributes_before_function(self, function=None, context=None):
super()._instantiate_attributes_before_function(function=function, context=context)
# create transient Parameters objects for custom function params
# done here because they need to be present before _instantiate_value which calls self.function
for param_name in self.cust_fct_params:
p = Parameter(self.cust_fct_params[param_name], modulable=True)
setattr(self.parameters, param_name, p)
p._set(p.default_value, context, skip_history=True)
class Parameters(Composition.Parameters):
""""""
optimizer = None
learning_rate = Parameter(.001, fallback_default=True)
losses = Parameter([])
trial_losses = Parameter([])
tracked_loss = Parameter(None, pnl_internal=True)
tracked_loss_count = Parameter(0, pnl_internal=True)
pytorch_representation = None
class Parameters(DistributionFunction.Parameters):
"""
Attributes
----------
bias
see `bias <DriftDiffusionAnalytical.bias>`
:default value: 0.5
:type: float
:read only: True
drift_rate
see `drift_rate <DriftDiffusionAnalytical.drift_rate>`
:default value: 1.0
:type: float
noise
see `noise <DriftDiffusionAnalytical.noise>`
:default value: 0.5
:type: float
starting_point
see `starting_point <DriftDiffusionAnalytical.starting_point>`
:default value: 0.0
:type: float
t0
see `t0 <DriftDiffusionAnalytical.t0>`
:default value: 0.2
:type: float
threshold
see `threshold <DriftDiffusionAnalytical.threshold>`
:default value: 1.0
:type: float
"""
drift_rate = Parameter(1.0,
modulable=True,
aliases=[MULTIPLICATIVE_PARAM])
starting_point = Parameter(0.0,
modulable=True,
aliases=[ADDITIVE_PARAM])
threshold = Parameter(1.0, modulable=True)
noise = Parameter(0.5, modulable=True)
t0 = .200
bias = Parameter(0.5,
read_only=True,
getter=_DriftDiffusionAnalytical_bias_getter)
class Parameters(ObjectiveMechanism.Parameters):
"""
Attributes
----------
variable
see `variable <ComparatorMechanism.variable>`
:default value: numpy.array([[0], [0]])
:type: ``numpy.ndarray``
:read only: True
function
see `function <ComparatorMechanism.function>`
:default value: `LinearCombination`(weights=numpy.array([[-1], [ 1]]))
:type: `Function`
output_ports
see `output_ports <ComparatorMechanism.output_ports>`
:default value: [`OUTCOME`]
:type: ``list``
:read only: True
sample
see `sample <ComparatorMechanism.sample>`
:default value: None
:type:
target
see `target <ComparatorMechanism.target>`
:default value: None
:type:
"""
# By default, ComparatorMechanism compares two 1D np.array input_ports
variable = Parameter(np.array([[0], [0]]),
read_only=True,
pnl_internal=True,
constructor_argument='default_variable')
function = Parameter(LinearCombination(weights=[[-1], [1]]),
stateful=False,
loggable=False)
sample = None
target = None
output_ports = Parameter(
[OUTCOME],
stateful=False,
loggable=False,
read_only=True,
structural=True,
)
class Parameters(TransferMechanism.Parameters):
"""
Attributes
----------
enable_learning
see `enable_learning <KohonenMechanism.enable_learning>`
:default value: True
:type: ``bool``
learning_function
see `learning_function <KohonenMechanism.learning_function>`
:default value: `Kohonen`
:type: `Function`
learning_rate
see `learning_rate <KohonenMechanism.learning_rate>`
:default value: None
:type:
matrix
see `matrix <KohonenMechanism.matrix>`
:default value: `AUTO_ASSIGN_MATRIX`
:type: ``str``
output_ports
see `output_ports <KohonenMechanism.output_ports>`
:default value: [`RESULT`, "{name: INPUT_PATTERN, variable: OWNER_VARIABLE}"]
:type: ``list``
:read only: True
"""
learning_function = Parameter(Kohonen(distance_function=GAUSSIAN),
stateful=False,
loggable=False,
reference=True)
learning_rate = Parameter(None, modulable=True)
enable_learning = True
matrix = DEFAULT_MATRIX
output_ports = Parameter(
[RESULT, {
NAME: INPUT_PATTERN,
VARIABLE: OWNER_VARIABLE
}],
stateful=False,
loggable=False,
read_only=True,
structural=True,
)
class Parameters(RecurrentTransferMechanism.Parameters):
"""
Attributes
----------
competition
see `competition <LCAMechanism.competition>`
:default value: 1.0
:type: float
initial_value
see `initial_value <LCAMechanism.initial_value>`
:default value: None
:type:
integrator_mode
see `integrator_mode <LCAMechanism.integrator_mode>`
:default value: True
:type: bool
leak
see `leak <LCAMechanism.leak>`
:default value: 0.5
:type: float
self_excitation
see `self_excitation <LCAMechanism.self_excitation>`
:default value: 0.0
:type: float
time_step_size
see `time_step_size <LCAMechanism.time_step_size>`
:default value: 0.1
:type: float
"""
function = Parameter(Logistic, stateful=False, loggable=False)
leak = Parameter(0.5, modulable=True)
auto = Parameter(0.0, modulable=True, aliases='self_excitation')
hetero = Parameter(-1.0, modulable=True)
competition = Parameter(1.0, modulable=True)
time_step_size = Parameter(0.1, modulable=True)
initial_value = None
integrator_mode = Parameter(True, setter=_integrator_mode_setter)
integrator_function = Parameter(LeakyCompetingIntegrator,
stateful=False,
loggable=False)
termination_measure = Parameter(max, stateful=False, loggable=False)
class Parameters(LearningMechanism.Parameters):
"""
Attributes
----------
function
see `function <KohonenLearningMechanism.function>`
:default value: `Hebbian`
:type: `Function`
learning_rate
see `learning_rate <KohonenLearningMechanism.learning_rate>`
:default value: None
:type:
learning_timing
see `learning_timing <KohonenLearningMechanism.learning_timing>`
:default value: LearningTiming.EXECUTION_PHASE
:type: `LearningTiming`
learning_type
see `learning_type <KohonenLearningMechanism.learning_type>`
:default value: LearningType.UNSUPERVISED
:type: `LearningType`
matrix
see `matrix <KohonenLearningMechanism.matrix>`
:default value: None
:type:
modulation
see `modulation <KohonenLearningMechanism.modulation>`
:default value: ModulationParam.ADDITIVE
:type: `ModulationParam`
"""
function = Parameter(Hebbian, stateful=False, loggable=False)
matrix = Parameter(None, modulable=True)
learning_rate = Parameter(None, modulable=True)
learning_type = LearningType.UNSUPERVISED
learning_timing = LearningTiming.EXECUTION_PHASE
modulation = ModulationParam.ADDITIVE
class Parameters(RecurrentTransferMechanism.Parameters):
"""
Attributes
----------
average_based
see `average_based <KWTAMechanism.average_based>`
:default value: False
:type: bool
function
see `function <KWTAMechanism.function>`
:default value: `Logistic`
:type: `Function`
inhibition_only
see `inhibition_only <KWTAMechanism.inhibition_only>`
:default value: True
:type: bool
k_value
see `k_value <KWTAMechanism.k_value>`
:default value: 0.5
:type: float
ratio
see `ratio <KWTAMechanism.ratio>`
:default value: 0.5
:type: float
threshold
see `threshold <KWTAMechanism.threshold>`
:default value: 0.0
:type: float
"""
function = Parameter(Logistic, stateful=False, loggable=False)
k_value = Parameter(0.5, modulable=True)
threshold = Parameter(0.0, modulable=True)
ratio = Parameter(0.5, modulable=True)
average_based = False
inhibition_only = True
class Parameters(LearningMechanism.Parameters):
"""
Attributes
----------
function
see `function <AutoAssociativeLearningMechanism.function>`
:default value: `Hebbian`
:type: `Function`
input_ports
see `input_ports <AutoAssociativeLearningMechanism.input_ports>`
:default value: [`ACTIVATION_INPUT`]
:type: ``list``
:read only: True
output_ports
see `output_ports <AutoAssociativeLearningMechanism.output_ports>`
:default value: ["{name: LearningSignal, variable: (OWNER_VALUE, 0)}"]
:type: ``list``
:read only: True
"""
function = Parameter(Hebbian, stateful=False, loggable=False)
modulation = ADDITIVE
input_ports = Parameter(
[ACTIVATION_INPUT],
stateful=False,
loggable=False,
read_only=True,
structural=True,
parse_spec=True,
)
output_ports = Parameter(
[{
NAME:
LEARNING_SIGNAL, # NOTE: This is the default, but is overridden by any LearningSignal arg
VARIABLE: (OWNER_VALUE, 0)
}],
stateful=False,
loggable=False,
read_only=True,
structural=True,
)
learning_type = LearningType.UNSUPERVISED
learning_timing = LearningTiming.EXECUTION_PHASE
class Parameters(CompositionFunctionApproximator.Parameters):
"""
Attributes
----------
prediction_vector
see `prediction_vector <RegressionCFA.prediction_vector>`
:default value: None
:type:
previous_state
see `previous_state <RegressionCFA.previous_state>`
:default value: None
:type:
regression_weights
see `regression_weights <RegressionCFA.regression_weights>`
:default value: None
:type:
update_weights
see `update_weights <RegressionCFA.update_weights>`
:default value: `BayesGLM`
:type: `Function`
"""
update_weights = Parameter(BayesGLM, stateful=False, loggable=False)
prediction_vector = None
previous_state = None
regression_weights = None
class Parameters(Function_Base.Parameters):
"""
Attributes
----------
filter_function
see `filter_function <PredictionMechanism.filter_function>`
:default value: None
:type:
rate
see `rate <PredictionMechanism.rate>`
:default value: 1.0
:type: float
window_size
see `window_size <PredictionMechanism.window_size>`
:default value: 1
:type: int
"""
variable = Parameter(None,
read_only=True,
pnl_internal=True,
constructor_argument='default_variable')
class Parameters(InterfaceFunction.Parameters):
corresponding_input_port = Parameter(
None,
structural=True,
stateful=False,
loggable=False
)
