For the complete documentation index, see llms.txt. Markdown versions of all pages are available by appending .md to any URL (e.g. /max/get-started.md).
Mojo struct
WithMaskFlashAttentionSplitKVCPU
struct WithMaskFlashAttentionSplitKVCPU
Implemented traits
AnyType,
ImplicitlyDestructible
Methods
execute
static def execute[dtype: DType, rank: Int, target: StringSlice[StaticConstantOrigin]](output: ManagedTensorSlice[Output, static_spec=output.static_spec], q: ManagedTensorSlice[Input, static_spec=q.static_spec], k: ManagedTensorSlice[FusedInput, static_spec=k.static_spec], v: ManagedTensorSlice[FusedInput, static_spec=v.static_spec], k_cache: ManagedTensorSlice[FusedInput, static_spec=k_cache.static_spec], v_cache: ManagedTensorSlice[FusedInput, static_spec=v_cache.static_spec], mask: ManagedTensorSlice[FusedInput, static_spec=mask.static_spec], scale: Float32, ctx: DeviceContext)
shape
static def shape[dtype: DType, rank: Int](q: ManagedTensorSlice[Input, static_spec=q.static_spec], k: ManagedTensorSlice[Input, static_spec=k.static_spec], v: ManagedTensorSlice[Input, static_spec=v.static_spec], k_cache: ManagedTensorSlice[Input, static_spec=k_cache.static_spec], v_cache: ManagedTensorSlice[Input, static_spec=v_cache.static_spec], mask: ManagedTensorSlice[Input, static_spec=mask.static_spec], scale: Float32) -> IndexList[rank]
Returns:
IndexList[rank]
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!