Xenova HF Staff

sync 2e7068faf55e

8d4a681 verified 5 days ago

5.1 kB

library_name: kernels
license: apache-2.0
tags:
  - kernel
  - webgpu
  - wgsl

com.microsoft.MatMulNBitsMlp

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

Description

Fuses a gated MLP over two block-quantized projections that share one activation: Y = silu(A_norm @ gate + gate_bias) * (A_norm @ up + up_bias), using the MatMulNBits weight packing with no zero-point input. A_norm is A, SimplifiedLayerNormalization(A, norm_scale), or SkipSimplifiedLayerNormalization(A, skip, norm_scale), whose residual sum may be returned as a second output. Only silu and the default accuracy_level = 0 are implemented; bfloat16 is not implemented.

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

Inputs

Name	Bind key	Logical dtype	Rank	Shape	Description	Presence
`A`	`aT`	`T1`	—	—	Shared activation of rank 2 `(M, K)` or rank 3 `(batch, sequence, K)`; only the last axis is the reduction axis.	required
`skip`	`skipT`	`T1`	—	—	Residual added to `A` before normalization, with `A`'s shape. Requires `norm_scale`.	optional
`norm_scale`	`normScaleT`	`T1`	`1`	—	Simplified-layer-normalization (RMS) gain of shape `[K]`. Absent means the projections read `A` unnormalized.	optional
`gate_B`	`gateBT`	`uint8`	`3`	—	Bit-packed uint8 gate weights of shape `(N, k_blocks, blob_size)`.	required
`gate_scales`	`gateScalesT`	`T1`	`2`	—	Per-block gate scales of shape `(N, k_blocks)`, with the same dtype as `A`. Quantization is symmetric: this operator has no zero-point input, so codes are offset by the midpoint `2^(bits - 1)`.	required
`gate_bias`	`gateBiasT`	`T1`	`1`	—	Optional gate bias of shape `[N]`, added before the activation.	optional
`up_B`	`upBT`	`uint8`	`3`	—	Bit-packed up weights, same shape and packing as gate_B.	required
`up_scales`	`upScalesT`	`T1`	`2`	—	Per-block up scales of shape `(N, k_blocks)`.	required
`up_bias`	`upBiasT`	`T1`	`1`	—	Optional up bias of shape `[N]`, added before the product.	optional

Outputs

Name	Bind key	Logical dtype	Rank	Shape	Description	Presence
`Y`	`yT`	`T1`	same as `A`	derived; see description	Gated MLP output: A's leading axes with a trailing N.	required
`input_skip_bias_sum`	`residualT`	`T1`	same as `A`	same as `A`	The residual sum A + skip, with A's shape. Requires the skip input.	optional

Attributes

Attributes and default values (overridable per request):

Attribute	Default	Description
`accuracy_level`	`0`	Minimum internal accuracy level: 0 (unset), 1 (float32), 2 (float16), 3 (bfloat16), or 4 (int8).
`bits`	`4`	Bit width used to quantize both weight matrices; this implementation supports 2, 4, and 8.
`epsilon`	`0.00001`	Epsilon used by the optional fused RMS normalization.
`K`	—	Input feature dimension shared by both quantized weight matrices.
`N`	—	Output feature dimension shared by both quantized weight matrices.
`activation`	—	Activation applied to the gate projection; this implementation supports `silu`.
`block_size`	—	Size of each quantization block along K.

Type constraints

Variable	Allowed dtypes
`T1`	`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
matmul-nbits-fused-rms-norm.wgsl.jinja
mlp-gate-up.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.MatMulNBitsMlp", { version: 1 });
const { yT } = await kernel({
  aT: { data: aTData, shape: [2, 16] },
  gateBT: { data: gateBTData, shape: [4, 2, 4] },
  gateScalesT: { data: gateScalesTData, shape: [4, 2] },
  upBT: { data: upBTData, shape: [4, 2, 4] },
  upScalesT: { data: upScalesTData, shape: [4, 2] },
}, {
  attrs: {
    K: 16,
    N: 4,
    block_size: 8,
    activation: "silu",
  },
});