ai.onnx.Gemm

ai.onnx · standard ONNX operator · ONNX opset ≥ 13

Description

Computes Y = alpha * A' * B' + beta * C, where A' and B' are optionally transposed views of the 2-D input matrices A and B, and C is an optional bias broadcastable to shape (M, N). Output Y has shape (M, N).

See the ONNX Gemm spec for the reference semantics.

Inputs

Name	Bind key	Logical dtype	Rank	Shape	Description	Presence
`A`	`a`	`T`	`2`	—	Input matrix `A` with shape `(M, K)` when `transA` is 0, or `(K, M)` when `transA` is non-zero.	required
`B`	`b`	`T`	`2`	—	Input matrix `B` with shape `(K, N)` when `transB` is 0, or `(N, K)` when `transB` is non-zero.	required
`C`	`bias`	`T`	—	—	Optional bias tensor broadcastable to `(M, N)`; treated as scalar 0 if absent.	optional

Outputs

Name	Bind key	Logical dtype	Rank	Shape	Description	Presence
`Y`	`y`	`T`	`2`	derived; see description	Output matrix of shape `(M, N)`.	required

Runtime arguments

Name	Kind	Semantic	Description	Presence
`M`	`u32`	`M`	Optional output-row count for the packed model-layout route; inferred from `A` for the standard matrix route.	optional
`inFeatures`	`u32`	`K`	Optional reduction dimension for the packed model-layout route; inferred from `A` and `B` for the standard matrix route.	optional
`outFeatures`	`u32`	`N`	Optional output-column count for the packed model-layout route; inferred from `B` for the standard matrix route.	optional

Attributes

Default values (overridable per request):

Attribute	Default	Description
`alpha`	`1`	Scalar multiplier applied to the matrix product A' * B'.
`beta`	`1`	Scalar multiplier applied to the optional bias tensor C.
`transA`	`0`	When non-zero, A is transposed before the multiplication.
`transB`	`0`	When non-zero, B is transposed before the multiplication.

Type constraints

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

Device requirements

Some implementation variants require subgroup-matrix, shader-f16, and subgroups. These are route-specific capabilities, not package-wide requirements; availability also depends on the request shape and dtype.

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
dense-tiled-matmul.wgsl.jinja
gemm-dense-f32-tiled-reg.wgsl.jinja
gemm-generic-tiled-reg.wgsl.jinja
gemm-model-m1-vec4.wgsl.jinja
gemm-subgroup-matrix.wgsl.jinja
gemm-transa-tail-correction.wgsl.jinja
gemm.model.wgsl.jinja
gemm.wgsl.jinja
matmul-notrans-vec4-tiled-reg.wgsl.jinja
reduce-axis0-splitk-combine.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/ai.onnx.Gemm", { version: 1 });
const { y } = await kernel({ a: { data: aData, shape: [2, 4] }, b: { data: bData, shape: [4, 3] } });

Downloads last month: -

kernel

webgpu

wgsl

apache-2.0

WebGPU

Requires WebGPU support. See the compatibility table.