com.microsoft.SkipLayerNormalization

com.microsoft · ONNX Runtime contrib operator · contrib since_version 1

Description

Fuses skip addition with layer normalization. The rank-3 standard surface currently supports float32, required beta, no bias or residual output, hidden sizes divisible by four, and exact or documented broadcast skip shapes. The provider's rank-2 extension supports float32 output-only with optional beta, or beta with optional bias when emitting the residual; its float16 path requires beta, bias, a residual output, and four-wide hidden size. Other combinations, bfloat16, and training statistics are not implemented.

See the ONNX Runtime SkipLayerNormalization contrib-operator spec for the reference semantics.

Inputs

Name	Bind key	Logical dtype	Rank	Shape	Description	Presence
`input`	`inputT`	`T`	—	—	Primary input normalized over the final hidden-size axis. Rank 3 is the public shape; rank 2 is an ONNX Runtime provider extension supported by this package.	required
`skip`	`skipT`	`T`	—	—	Residual tensor. For rank-3 input it is exact shape, `(1, sequence_length, hidden_size)`, or `(sequence_length, hidden_size)`; rank-2 input requires exact shape.	required
`gamma`	`gammaT`	`T`	`1`	—	Layer-norm scale weights of shape `(hidden_size)`.	required
`beta`	`betaT`	`T`	`1`	—	Layer-norm bias weights of shape `(hidden_size)`.	optional
`bias`	`biasT`	`T`	`1`	—	Optional additive bias of shape `(hidden_size)` added to `input + skip` before normalization.	optional

Outputs

Name	Bind key	Logical dtype	Rank	Shape	Description	Presence
`output`	`outputT`	`T`	same as `input`	same as `input`	Normalized output tensor with the same shape as `input`.	required
`input_skip_bias_sum`	`residualT`	`T`	same as `input`	same as `input`	Sum of `input`, `skip`, and `bias` (when present) before normalization, with the same shape as `input`.	optional

Attributes

Default values (overridable per request):

Attribute	Default	Description
`epsilon`	`9.999999960041972e-13`	Non-negative epsilon added to the variance before taking the square root.

Type constraints

Variable	Allowed dtypes
`T`	`float32`, `float16`

Files

metadata.json — kernel metadata (id, digests, provenance)
manifest.json — the op contract (source of truth)
test.json — correctness cases
bench.json — benchmark + tuning cases
norm-skip-row-vec4.wgsl.jinja
norm-skip-row.wgsl.jinja

Use with `@huggingface/kernels`

The loader derives every required output's shape and logical dtype from the manifest contract and this call. It then allocates the result tensors automatically.

The version: 1 option selects the published kernel contract; it is independent of any operator opset, contrib since_version, or model version.

Replace each *Data placeholder with a typed array containing the corresponding input data.

import { getKernel } from "@huggingface/kernels";

const kernel = await getKernel("webgpu-kernels/com.microsoft.SkipLayerNormalization", { version: 1 });
const { outputT } = await kernel({
  inputT: { data: inputTData, shape: [2, 4] },
  skipT: { data: skipTData, shape: [2, 4] },
  gammaT: { data: gammaTData, shape: [4] },
});

Downloads last month: -

kernel

webgpu

wgsl

apache-2.0

WebGPU

Requires WebGPU support. See the compatibility table.