superneuroabm package

Subpackages

• superneuroabm.core package
  • Submodules
  • superneuroabm.core.agent module
    • AgentFactory
      • AgentFactory.breeds
      • AgentFactory.create_agent()
      • AgentFactory.generate_agent_data_tensors()
      • AgentFactory.get_agent_property_value()
      • AgentFactory.get_agents_with()
      • AgentFactory.num_agents
      • AgentFactory.register_breed()
      • AgentFactory.set_agent_property_value()
      • AgentFactory.update_agents_properties()
    • Breed
      • Breed.name
      • Breed.properties
      • Breed.register_property()
      • Breed.register_step_func()
      • Breed.step_funcs
  • superneuroabm.core.model module
    • Model
      • Model.THREADSPERBLOCK
      • Model.create_agent_of_breed()
      • Model.get_agent_property_value()
      • Model.get_agents_with()
      • Model.get_synapse_weight()
      • Model.load()
      • Model.register_breed()
      • Model.save()
      • Model.set_agent_property_value()
      • Model.setup()
      • Model.simulate()
    • generate_gpu_func()
    • smap()
  • superneuroabm.core.util module
    • compress_tensor()
    • convert_to_equal_side_tensor()
  • Module contents

Submodules

superneuroabm.model module

Model class for building an SNN

class superneuroabm.model.NeuromorphicModel(neuron_breed_info: ~typing.Dict[str, ~typing.List[~typing.Callable]] = {'Neuron': [<function neuron_step_func>, <function synapse_step_func>]})

Bases: Model

add_spike(neuron_id: int, tick: int, value: float) → None

Schedules an external input spike to this Neuron.

Parameters:

• tick – tick at which spike should be triggered

• value – spike value

create_neuron(breed: str = 'Neuron', threshold: float = 1, reset_state: float = 0, leak: float = 0, refractory_period: int = 0, axonal_delay: int = 1) → int

Creates and Neuron agent.

Returns:

SAGESim agent id of neuron

create_synapse(pre_neuron_id: int, post_neuron_id: int, weight: int = 1, synaptic_delay: int = 1, synapse_learning_params: List[float] | None = None) → None

Creates and adds Synapse agent.

Parameters:

• pre_neuron_id – int presynaptic neuron id

• post_neuron_id – int postsynaptic neuron id

• weight – weight of synapse

• delay (synaptic) – number of timesteps to delay synapse by

• synapse_learning_params – Optional. Any parameters used in a learning enabled step function. Must be specified in order of use in step function.

get_spike_times(neuron_id: int) → array

load(fpath: str)

Loads model from pickle file.

Parameters:

fpath – file path to pickle file.

save(fpath: str)

Saves model. Must be overridden if additional data pertaining to application must be saved.

Parameters:

• fpath – file path to save pickle file at

• app_data – additional application data to be saved.

setup(use_cuda: bool = False, output_buffer_len: int = 1000, retain_weights=False) → None

Resets the simulation and initializes agents.

Parameters:

retain_weights – False by default. If True, updated weights are not reset upon setup.

simulate(ticks: int, update_data_ticks: int = 1, num_cpu_proc: int = 4) → None

Override of superneuroabm.core.model mainly to register an AgentDataCollector to monitor marked output Neurons.

summary() → str

Verbose summary of the network structure.

Returns:

str information of netowkr struture

update_synapse(pre_neuron_id: int, post_neuron_id: int, weight: int | None = None, synapse_learning_params: List[float] | None = None)

superneuroabm.neuron module

Neuron and synapse step functions for spiking neural networks

superneuroabm.neuron.neuron_step_func(global_data_vector, breeds, thresholds, reset_states, leaks, refractory_periods, output_synapsess, t_elapses, internal_state, neuron_delay_regs, input_spikess, output_synapses_learning_paramss, output_spikess, my_idx)

superneuroabm.neuron.synapse_step_func(global_data_vector, breeds, thresholds, reset_states, leaks, refractory_periods, output_synapsess, t_elapses, internal_state, neuron_delay_regs, input_spikess, output_synapses_learning_paramss, output_spikess, my_idx)

superneuroabm.neuron.synapse_with_stdp_step_func(global_data_vector, breeds, thresholds, reset_states, leaks, refractory_periods, output_synapsess, t_elapses, internal_state, neuron_delay_regs, input_spikess, output_synapses_learning_paramss, output_spikess, my_idx)

Module contents