MCPcopy Create free account
hub / github.com/pytorch/executorch / SpecAllocResult

Class SpecAllocResult

exir/memory_planning.py:730–739  ·  view source on GitHub ↗

These are the values that a memory plannig algorithm assigns to a spec. These are not directly written back into the spec object, but are used to track the allocation decisions and assigned back to the spec object in the end, based on which algorithm is picked as the best performing one.

Source from the content-addressed store, hash-verified

728
729@dataclass
730class SpecAllocResult:
731 """These are the values that a memory plannig algorithm assigns to a spec.
732 These are not directly written back into the spec object, but are used to
733 track the allocation decisions and assigned back to the spec object in the
734 end, based on which algorithm is picked as the best performing one.
735 """
736
737 mem_id: int
738 mem_obj_id: int
739 mem_offset: int
740
741
742@dataclass

Callers 2

greedyFunction · 0.85
naiveFunction · 0.85

Calls

no outgoing calls

Tested by

no test coverage detected