Interface: LtnPredicateFit
@kortexya/reasoninglayer / LtnPredicateFit
Interface: LtnPredicateFit
Section titled “Interface: LtnPredicateFit”Defined in: src/types/ltn.ts:405
The outcome of a neural-predicate fit.
Properties
Section titled “Properties”examples
Section titled “examples”examples:
number
Defined in: src/types/ltn.ts:411
Number of examples trained on.
finalLoss
Section titled “finalLoss”finalLoss:
number
Defined in: src/types/ltn.ts:409
Final-epoch mean training loss (MSE).
inputDim
Section titled “inputDim”inputDim:
number
Defined in: src/types/ltn.ts:413
The predicate’s input dimension.
persisted
Section titled “persisted”persisted:
boolean
Defined in: src/types/ltn.ts:418
true when the weights were checkpointed to the per-tenant store (an
LTN_WEIGHTS_DIR is configured); false when persistence is off.
predicate
Section titled “predicate”predicate:
string
Defined in: src/types/ltn.ts:407
The predicate name fitted.