jaxsnn.event.from_nir_data

Translate NIRGraphData to a jaxsnn-compatible spike representation.

Classes

EventPropSpike(time, idx, current)
NIRGraphData(nodes, …)	Dictionary of NIRNodeData where each entry represents a NIRNode of a corresponding NIRGraph with its observables.
NIRNodeData(observables)	Dictionary of EventData or TimeGriddedData where each entry represents an observable (e.g., spikes, voltages) of a corresponding NIRNode
TimeGriddedData(data, dt)	Time gridded data of shape (n_samples, n_time_steps, n_neurons) with binary entries (0 or 1) indicating whether a neuron spiked at a particular time step dt.

Functions

jaxsnn.event.from_nir_data.from_nir_data(nir_graph_data: nir.data_ir.graph.NIRGraphData, jaxsnn_model) → Dict[str, jaxsnn.event.types.EventPropSpike]

Convert NIRGraphData to a dict of EventPropSpikes (jax-snn representation)

A linear noise is added on the spike times if the incoming data is time-gridded.