File size: 2,326 Bytes
a31d703 fc9d701 a31d703 fc9d701 a31d703 fc9d701 6c10a5f fc9d701 6c10a5f fc9d701 6c10a5f fc9d701 6c10a5f fc9d701 6c10a5f fc9d701 6c10a5f fc9d701 6c10a5f fc9d701 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 | ---
library_name: kernels
license: apache-2.0
tags:
- kernel
- webgpu
- wgsl
---
# ai.onnx.Exp
`ai.onnx` · standard ONNX operator · ONNX opset ≥ 13
## Description
Computes `exp(x)` elementwise for every element of the input tensor. The output has the same shape and type as the input.
See the [ONNX `Exp` spec](https://onnx.ai/onnx/operators/onnx__Exp.html) for the reference semantics.
## Inputs
| Name | Upstream name | Logical dtype | Rank | Shape | Description | Presence |
| --- | --- | --- | --- | --- | --- | --- |
| `x` | `input` | `T` | — | — | Values used as exponents of `e` in the elementwise exponential. | required |
## Outputs
| Name | Upstream name | Logical dtype | Rank | Shape | Description | Presence |
| --- | --- | --- | --- | --- | --- | --- |
| `y` | `output` | `T` | same as `x` | same as `x` | Output tensor containing the elementwise exponential of the input. | required |
## Type constraints
| Variable | Allowed dtypes |
| --- | --- |
| `T` | `float32`, `float16` |
## Files
- [`metadata.json`](build/webgpu/metadata.json) — kernel metadata (id, digests, per-variant templates, 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
- [`unary-scalar.wgsl.jinja`](build/webgpu/unary-scalar.wgsl.jinja)
- [`unary-vec4.wgsl.jinja`](build/webgpu/unary-vec4.wgsl.jinja)
## Use with `@huggingface/kernels`
```sh
npm install --save-exact @huggingface/kernels@0.0.1-preview.2
```
Required output shapes and logical data types are inferred from the supplied inputs and attributes; result tensors are allocated automatically.
The `version: 1` option selects the published kernel contract; it is independent of any operator opset, contrib `since_version`, or model version.
It follows the `v1` branch as fixes land. To pin exact artifact bytes, pass a 40-character commit `revision` instead of `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/ai.onnx.Exp", { version: 1 });
const { y } = await kernel({ x: { data: xData, shape: [] } });
```
|