MixedResult#
- class hybrid_aif.planning.MixedResult[source]#
Bases:
NamedTupleOne mixed plan, over
Pdiscrete policies, each with optimised continuous controls. With several starts,planis the best start’s for each discrete policy andtrace,gradient_traceandbest_stepare that start’s (startsis 1 for a single start).- q_pi: Categorical#
The posterior over the
Pdiscrete policies.
- tree: dict[str, dict[str, Interval]] | None#
Their term trees at the optimised plans (
Noneundermetrics=False).
- plan: Array#
Their optimised continuous controls,
(P, policy_len, action_dim)(the continuous nodes’ blocks concatenated in node order).
- best_step: Array#
The iterate returned,
(P,), as inGradientResult.
- static __new__(
- _cls,
- q_pi,
- tree,
- discrete,
- plan,
- G,
- trace,
- gradient_trace,
- best_step,
- start_scores,
- start_plans,
Create new instance of MixedResult(q_pi, tree, discrete, plan, G, trace, gradient_trace, best_step, start_scores, start_plans)