ParameterSlotRole
Enums · AiDotNet.Models.Parameters
Describes how a numeric slot participates in model state and optimization.
Fields
| Field | Summary |
|---|---|
ParameterSlotRole Alias |
A second name for storage owned by another slot. |
ParameterSlotRole Buffer |
Persistent auxiliary state that is neither fitted nor optimizer-updated. |
ParameterSlotRole External |
State owned by an external runtime, such as a loaded ONNX graph. |
ParameterSlotRole Frozen |
A restorable value intentionally excluded from optimization. |
ParameterSlotRole Gradient |
A gradient accumulator rather than model state. |
ParameterSlotRole InputSizedState |
Fitted state whose extent comes from the caller’s data, so it is restored by name but never joins the flat parameter vector. |
ParameterSlotRole LearnedState |
Fitted state that is restored but is not updated by a gradient optimizer. |
ParameterSlotRole Scratch |
Transient working storage that is neither restored nor optimized. |
ParameterSlotRole Trainable |
A value an optimizer is allowed to update. |
int value__ |