ModelSplitMode enum
Strategy for distributing model tensors across GPU devices.
Mirrors llama.cpp llama_split_mode. The default, layer, preserves
upstream behavior.
Values
- none → const ModelSplitMode
-
Use a single GPU selected by ModelParams.mainGpu.
const ModelSplitMode(0) - layer → const ModelSplitMode
-
Split layers and KV cache across GPUs.
const ModelSplitMode(1) - row → const ModelSplitMode
-
Split layers and KV cache across GPUs, with row-level tensor splitting where supported by the backend.
const ModelSplitMode(2) - tensor → const ModelSplitMode
-
Use tensor parallelism where supported by the backend and model.
const ModelSplitMode(3)
Properties
- hashCode → int
-
The hash code for this object.
no setterinherited
- index → int
-
A numeric identifier for the enumerated value.
no setterinherited
- llamaCppValue → int
-
Native llama.cpp enum value.
final
- name → String
-
Available on Enum, provided by the EnumName extension
The name of the enum value.no setter - runtimeType → Type
-
A representation of the runtime type of the object.
no setterinherited
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
Constants
-
values
→ const List<
ModelSplitMode> - A constant List of the values in this enum, in order of their declaration.