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
ReduceRMSNormFusedResidualAdd
struct ReduceRMSNormFusedResidualAdd
Implemented traits
AnyType,
ImplicitlyDestructible
Methods
execute
static def execute[dtype: DType, rank: Int, target: StringSlice[StaticConstantOrigin], multiply_before_cast: Bool = True](output: ManagedTensorSlice[Output, static_spec=output.static_spec], residual_output: ManagedTensorSlice[Output, static_spec=residual_output.static_spec], input: ManagedTensorSlice[FusedInput, static_spec=input.static_spec], residual_input: ManagedTensorSlice[FusedInput, static_spec=residual_input.static_spec], gamma1: ManagedTensorSlice[Input, static_spec=gamma1.static_spec], gamma2: ManagedTensorSlice[Input, static_spec=gamma2.static_spec], epsilon1: Scalar[dtype], epsilon2: Scalar[dtype], weight_offset1: Scalar[dtype], weight_offset2: Scalar[dtype], ctx: DeviceContext)
shape
static def shape[dtype: DType, rank: Int](input: ManagedTensorSlice[Input, static_spec=input.static_spec], residual_input: ManagedTensorSlice[Input, static_spec=residual_input.static_spec], gamma1: ManagedTensorSlice[Input, static_spec=gamma1.static_spec], gamma2: ManagedTensorSlice[Input, static_spec=gamma2.static_spec], epsilon1: Scalar[dtype], epsilon2: Scalar[dtype], weight_offset1: Scalar[dtype], weight_offset2: Scalar[dtype]) -> IndexList[rank]
Returns:
Was this page helpful?
Thank you! We'll create more content like this.
Thank you for helping us improve!