Loihi Parameter Tuning Dashboard

This tool lets you change some of the main parameters on Loihi via an interactive dashboard. It is built on top of the Brian2Loihi Emulator (https://github.com/sagacitysite/brian2_loihi).

Requirements

• numpy
• ipywidgets
• matplotlib
• brian2 (2.4.2 or higher)
• brian2_loihi (0.5.2 or higher)

Available Parameters

The following parameters can be manipulated within the tool:

• runtime (integer, 0...200): defines how long the simulation is in unitless timesteps
• decay_v (integer, 0...4096): controls the steepness of the membrane voltage decay
• decay_I (integer, 0...4096): controls the steepness of the synaptic input decay
• weight_mantissa (integer, 0...255): mantissa part of synaptic weight
• weight_exponent (integer, -8...7): weight exponent which scales the weight mantissa by w_mant * 2^(6 + w_exp)
• numWeightBits (integer, 0...8): defines the precision of the weight, default is 8 bits
• threshold_v_mant (integer, 0...131071): The mantissa of the membrane voltage threshold, actual threshold is threshold_v_mant * 2^6
• ref_time (integer, 0...64): length of the refractory period
• scale (linear or log): toggles the scale of the voltage and input plots

Example

A photo of the dashboard is shown below. Run the jupyter notbook to move each of the sliders and see the neuron's behavior change in real time.