pyhgf.utils.vectorised_belief_propagation.learn_sweep#

pyhgf.utils.vectorised_belief_propagation.learn_sweep(network, opt_state, optimiser, learning_kind, synaptic_uncertainty_settings=None)[source]#

JIT-compiled weight-learning phase.

See _learn_sweep().

Parameters:

  • network (VectorisedNetwork)

  • opt_state (optax.OptState)

  • optimiser (Optional[optax.GradientTransformation])

  • learning_kind (str)

  • synaptic_uncertainty_settings (Optional[SynapticUncertaintySettings])

Return type:

tuple[VectorisedNetwork, optax.OptState]