MemoryRerankDecision Python API

A validated sparse selection and its portable inference metadata.

Attributes

attributeselected_rankstuple[int, ...]

attributeusageInferenceUsage

attributediscarded_rank_countint

= 0

attributeused_fallbackbool

= False

constructor__init__(selected_ranks, usage, discarded_rank_count=0, used_fallback=False) -> None

paramselected_rankstuple[int, ...]

paramusageInferenceUsage

paramdiscarded_rank_countint

= 0

paramused_fallbackbool

= False

Returns

None