| --- |
| library_name: kernels |
| license: apache-2.0 |
| tags: |
| - kernel |
| - webgpu |
| - wgsl |
| --- |
| # com.microsoft.FastGelu |
|
|
| `com.microsoft` · ONNX Runtime contrib operator · contrib since_version 1 |
| |
| ## Description |
| |
| Applies the GELU (Gaussian Error Linear Unit) activation using a tanh approximation: `Y = 0.5 * X * (1 + tanh(0.797885 * X + 0.035677 * X^3))`. An optional `bias` is added to `X` before the activation is computed. This WebGPU package implements float16 and float32; the schema-allowed double and bfloat16 types are not supported. |
| |
| See the [ONNX Runtime `FastGelu` contrib-operator spec](https://github.com/microsoft/onnxruntime/blob/main/docs/ContribOperators.md#com.microsoft.FastGelu) for the reference semantics. |
| |
| ## Inputs |
| |
| | Name | Bind key | Logical dtype | Rank | Shape | Description | Presence | |
| | --- | --- | --- | --- | --- | --- | --- | |
| | `X` | `X` | `T` | — | — | Values transformed by FastGelu after adding the optional `bias`. | required | |
| | `bias` | `bias` | `T` | `1` | — | Optional 1-D bias added to `X` along the last dimension before the GELU activation. | optional | |
| |
| ## Outputs |
| |
| | Name | Bind key | Logical dtype | Rank | Shape | Description | Presence | |
| | --- | --- | --- | --- | --- | --- | --- | |
| | `Y` | `Y` | `T` | same as `X` | same as `X` | Output tensor after applying the GELU activation; same shape as `X`. | required | |
| |
| ## Type constraints |
| |
| | Variable | Allowed dtypes | |
| | --- | --- | |
| | `T` | `float32`, `float16` | |
| |
| ## Files |
| |
| - [`metadata.json`](build/webgpu/metadata.json) — kernel metadata (id, digests, provenance) |
| - [`manifest.json`](build/webgpu/manifest.json) — the op contract (source of truth) |
| - [`test.json`](build/webgpu/test.json) — correctness cases |
| - [`bench.json`](build/webgpu/bench.json) — benchmark + tuning cases |
| - [`elementwise-bias-gelu.wgsl.jinja`](build/webgpu/elementwise-bias-gelu.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. |
|
|
| ```js |
| import { getKernel } from "@huggingface/kernels"; |
| |
| const kernel = await getKernel("webgpu-kernels/com.microsoft.FastGelu", { version: 1 }); |
| const { Y } = await kernel({ X: { data: XData, shape: [5] } }); |
| ``` |
|
|
