Planner#
- class hybrid_aif.planning.Planner[source]#
Bases:
ProtocolPolicy inference over a
PlannerContext.Returns
(q_pi, tree, policies): the policy posterior, the term tree (per-policy leaves aligned withq_pi), and the planner’s own policy set, int32(num_policies, policy_len, K)(one column per control node) or, over continuous actions, float(num_policies, policy_len, action_dim). Callers sample from these policies.metrics=FalsereturnsNonefor the tree and must leaveq_piunchanged up to float reassociation. A planner that builds no tree returns{}undermetrics=True.A planner that takes a warm start carries the mark
takes_init = Trueand accepts the keywordinit, the continuous plan(policy_len, action_dim)its search starts from (the gradient and mixed planners;planner_takes_init()).- __init__(*args, **kwargs)#