| --- |
| library_name: kernels |
| license: apache-2.0 |
| tags: |
| - kernel |
| - webgpu |
| - wgsl |
| --- |
| # ai.onnx.Range |
|
|
| `ai.onnx` · standard ONNX operator · ONNX opset ≥ 27 |
|
|
| ## Description |
|
|
| Generates a 1-D tensor of numbers starting at `start`, incrementing by `delta`, up to but not including `limit`. The output length is `max(ceil((limit - start) / delta), 0)`, and element `i` equals `start + i * delta`. Float16 inputs use float32 intermediate arithmetic when `stash_type` is `1`, then cast each result back to float16. |
|
|
| See the [ONNX `Range` spec](https://onnx.ai/onnx/operators/onnx__Range.html) for the reference semantics. |
|
|
| ## Inputs |
|
|
| | Name | Bind key | Logical dtype | Rank | Shape | Description | Presence | |
| | --- | --- | --- | --- | --- | --- | --- | |
| | `start` | `start` | `T` | `0` | — | Scalar first value in the output sequence. | required | |
| | `limit` | `limit` | `T` | `0` | — | Scalar exclusive upper bound of the sequence. | required | |
| | `delta` | `delta` | `T` | `0` | — | Scalar step size between consecutive output values. | required | |
|
|
| ## Outputs |
|
|
| | Name | Bind key | Logical dtype | Rank | Shape | Description | Presence | |
| | --- | --- | --- | --- | --- | --- | --- | |
| | `output` | `output` | `T` | `1` | — | 1-D tensor containing the generated range of values, with the same dtype as the inputs. | required | |
|
|
| ## Attributes |
|
|
| Default values (overridable per request): |
|
|
| | Attribute | Default | Description | |
| | --- | --- | --- | |
| | `stash_type` | `1` | TensorProto element type used for float16 intermediate arithmetic; this implementation supports the standard float32 value (`1`). It has no effect for float32 and int32 inputs. | |
|
|
| ## Type constraints |
|
|
| | Variable | Allowed dtypes | |
| | --- | --- | |
| | `T` | `float32`, `float16`, `int32` | |
|
|
| ## 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 |
| - [`range.wgsl.jinja`](build/webgpu/range.wgsl.jinja) |
|
|
| ## Use with `@huggingface/kernels` |
|
|
| The loader automatically allocates outputs whose metadata it can derive from the manifest contract and this call. |
|
|
| The explicit `outputs` entries provide shape and logical dtype metadata for the results listed below: |
|
|
| - `output` |
|
|
| Each entry either requests an optional result or supplies metadata that cannot be inferred from the inputs. |
|
|
| 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/ai.onnx.Range", { version: 1 }); |
| // Explicit destinations request optional results or supply metadata that cannot be inferred. |
| const { output } = await kernel({ |
| start: { data: startData, shape: [] }, |
| limit: { data: limitData, shape: [] }, |
| delta: { data: deltaData, shape: [] }, |
| }, { |
| outputs: { output: { shape: [1], dtype: "float32" } }, |
| }); |
| ``` |
|
|
