spark.nn.components.synapses

spark.nn.components.synapses#

Submodules#

Classes#

`Synanpses`	Abstract synapse model.
`SynanpsesOutput`	Generic synapses model output spec.
`LinearSynapses`	Linea synaptic model.
`LinearSynapsesConfig`	LinearSynapses model configuration class.
`TracedSynapses`	Traced synaptic model.
`TracedSynapsesConfig`	TracedSynapses model configuration class.
`RDTracedSynapses`	Rise-Decay traced synaptic model.
`RDTracedSynapsesConfig`	RDTracedSynapses model configuration class.
`RFSTracedSynapses`	Traced synaptic model.
`RFSTracedSynapsesConfig`	RFSTracedSynapses model configuration class.

Package Contents#

class spark.nn.components.synapses.Synanpses(config=None, **kwargs)[source]#

Bases: spark.nn.components.base.Component, Generic[ConfigT]

Abstract synapse model.

Note that we require the kernel entries to be in pA for numerical stability, since most of the time we want to run in half-precision. However somas expect the current in nA so we need to rescale the output.

Init:

Input:: spikes: SpikeArray
Output:: currents: CurrentArray

Parameters:: config (ConfigT | None)

abstractmethod get_kernel()[source]#

Return type:: spark.core.payloads.FloatArray

abstractmethod set_kernel(new_kernel)[source]#

Parameters:: new_kernel (spark.core.payloads.FloatArray)
Return type:: None

__call__(spikes)[source]#

Compute synanpse’s currents.

Parameters:: spikes (spark.core.payloads.SpikeArray)
Return type:: SynanpsesOutput

class spark.nn.components.synapses.SynanpsesOutput[source]#

Bases: TypedDict

Generic synapses model output spec.

Initialize self. See help(type(self)) for accurate signature.

currents: spark.core.payloads.CurrentArray[source]#

class spark.nn.components.synapses.LinearSynapses(config=None, **kwargs)[source]#

Bases: spark.nn.components.synapses.base.Synanpses

Linea synaptic model. Output currents are computed as the dot product of the kernel with the input spikes.

Init:: units: tuple[int, …] kernel: jax.Array | Initializer
Input:: spikes: SpikeArray
Output:: currents: CurrentArray
Reference:: Neuronal Dynamics: From Single Neurons to Networks and Models of Cognition. Gerstner W, Kistler WM, Naud R, Paninski L. Chapter 1.3 Integrate-And-Fire Models https://neuronaldynamics.epfl.ch/online/Ch1.S3.html

Parameters:: config (LinearSynapses | None)

config: LinearSynapsesConfig[source]#

build(input_specs)[source]#

Build method.

Parameters:: input_specs (dict[str, spark.core.specs.PortSpecs])

get_kernel()[source]#

Return type:: spark.core.payloads.FloatArray

get_flat_kernel()[source]#

Return type:: spark.core.payloads.FloatArray

set_kernel(new_kernel)[source]#

Parameters:: new_kernel (spark.core.payloads.FloatArray)
Return type:: None

class spark.nn.components.synapses.LinearSynapsesConfig(__skip_validation__=False, **kwargs)[source]#

Bases: spark.nn.components.synapses.base.SynanpsesConfig

LinearSynapses model configuration class.

Parameters:: __skip_validation__ (bool)

units: tuple[int, Ellipsis][source]#

kernel: jax.Array | spark.nn.initializers.base.Initializer[source]#

class spark.nn.components.synapses.TracedSynapses(config=None, **kwargs)[source]#

Bases: spark.nn.components.synapses.linear.LinearSynapses

Traced synaptic model. Output currents are computed as the trace of the dot product of the kernel with the input spikes.

Init:: units: tuple[int, …] kernel: KernelInitializerConfig tau: float | jax.Array scale: float | jax.Array base: float | jax.Array
Input:: spikes: SpikeArray
Output:: currents: CurrentArray

Parameters:: config (TracedSynapsesConfig | None)

config: TracedSynapsesConfig[source]#

current_tracer: spark.core.tracers.Tracer[source]#

build(input_specs)[source]#

Build method.

Parameters:: input_specs (dict[str, spark.core.specs.PortSpecs])

reset()[source]#

Resets component state.

Return type:: None

class spark.nn.components.synapses.TracedSynapsesConfig(__skip_validation__=False, **kwargs)[source]#

Bases: spark.nn.components.synapses.linear.LinearSynapsesConfig

TracedSynapses model configuration class.

Parameters:: __skip_validation__ (bool)

tau: float | jax.Array | spark.nn.initializers.base.Initializer[source]#

scale: float | jax.Array[source]#

base: float | jax.Array[source]#

class spark.nn.components.synapses.RDTracedSynapses(config=None, **kwargs)[source]#

Bases: spark.nn.components.synapses.linear.LinearSynapses

Rise-Decay traced synaptic model. Output currents are computed as the RDTrace of the dot product of the kernel with the input spikes.

Init:: units: tuple[int, …] kernel: KernelInitializerConfig tau_rise: float | jax.Array scale_rise: float | jax.Array base_rise: float | jax.Array tau_decay: float | jax.Array scale_decay: float | jax.Array base_decay: float | jax.Array
Input:: spikes: SpikeArray
Output:: currents: CurrentArray

Parameters:: config (RDTracedSynapsesConfig | None)

config: RDTracedSynapsesConfig[source]#

current_tracer: spark.core.tracers.RDTracer[source]#

build(input_specs)[source]#

Build method.

Parameters:: input_specs (dict[str, spark.core.specs.PortSpecs])

reset()[source]#

Resets component state.

Return type:: None

class spark.nn.components.synapses.RDTracedSynapsesConfig(__skip_validation__=False, **kwargs)[source]#

Bases: spark.nn.components.synapses.linear.LinearSynapsesConfig

RDTracedSynapses model configuration class.

Parameters:: __skip_validation__ (bool)

tau_rise: float | jax.Array | spark.nn.initializers.base.Initializer[source]#

scale_rise: float | jax.Array[source]#

base_rise: float | jax.Array[source]#

tau_decay: float | jax.Array | spark.nn.initializers.base.Initializer[source]#

scale_decay: float | jax.Array[source]#

base_decay: float | jax.Array[source]#

class spark.nn.components.synapses.RFSTracedSynapses(config=None, **kwargs)[source]#

Bases: spark.nn.components.synapses.linear.LinearSynapses

Traced synaptic model. Output currents are computed as the trace of the dot product of the kernel with the input spikes.

Init:: units: tuple[int, …] kernel: KernelInitializerConfig alpha: float | jax.Array tau_rise: float | jax.Array scale_rise: float | jax.Array base_rise: float | jax.Array tau_fast_decay: float | jax.Array scale_fast_decay: float | jax.Array base_fast_decay: float | jax.Array tau_slow_decay: float | jax.Array scale_slow_decay: float | jax.Array base_slow_decay: float | jax.Array
Input:: spikes: SpikeArray
Output:: currents: CurrentArray