Network

class brainpy.state.network.Network(*args, **kwargs)

brainpy.state network base class.

Subclass and define populations, projections, and devices as attributes. update() walks the immediate module-tree children in projection-first then dynamics order.

simulate(duration, *, dt=None, monitor=None)

Run the network for duration.

Wraps brainstate.transform.for_loop over self.update.

Parameters:

• duration (brainunit.Quantity) – Wall-clock time to simulate.

• dt (brainunit.Quantity, optional) – Timestep override. Defaults to brainstate.environ.get('dt').

• monitor (list[str] | dict[str, Callable] | None) –

  Per-step recording specification:

  • None (default) — no monitoring; the function returns {}.

  • list of dotted attribute paths (e.g. ['exc.spike']) — stacked across time and returned under that key.

  • dict of name -> callable(net): callable evaluated each step, result stacked under name.