LoadOptions class
Placement knobs applied when a model is loaded.
Constructors
-
LoadOptions({List<
BackendPreference> backendPreferences = const <BackendPreference>[], AcceleratorPolicy? accelerator, int? contextLength, int? threads, bool forceReload = false, InferenceFramework? framework, bool? useGpu}) -
Pin an engine or override runtime sizing for one load.
const
Properties
- accelerator → AcceleratorPolicy?
-
Hardware class to run on.
final
-
backendPreferences
→ List<
BackendPreference> -
Ranked backend choices; the first entry the platform can honor wins.
final
- contextLength → int?
-
Context window in tokens.
final
- forceReload → bool
-
Reload even when the model is already resident.
final
- framework → InferenceFramework?
-
Deprecated: use backendPreferences. Engine to load with.
final
- hashCode → int
-
The hash code for this object.
no setterinherited
- resolvedAccelerator → AcceleratorPolicy?
-
accelerator, folding in the deprecated useGpu alias.
no setter
-
resolvedBackendPreferences
→ List<
BackendPreference> -
backendPreferences, folding in the deprecated framework alias.
no setter
- runtimeType → Type
-
A representation of the runtime type of the object.
no setterinherited
- threads → int?
-
CPU threads the engine may use.
final
- useGpu → bool?
-
Deprecated: use accelerator. True forces GPU offload, false forbids
it, null lets the engine decide.
final
Methods
-
noSuchMethod(
Invocation invocation) → dynamic -
Invoked when a nonexistent method or property is accessed.
inherited
-
toString(
) → String -
A string representation of this object.
inherited
Operators
-
operator ==(
Object other) → bool -
The equality operator.
inherited