--- library_name: kernels license: apache-2.0 tags: - kernel - webgpu - wgsl --- # ai.onnx.Constant `ai.onnx` · standard ONNX operator · ONNX opset ≥ 24 ## Description Produces a numeric constant from ONNX's tensor-valued `value` attribute or its float scalar/list shorthands. Sparse, string, and int64-only shorthand attributes are not yet supported. See the [ONNX `Constant` spec](https://onnx.ai/onnx/operators/onnx__Constant.html) for the reference semantics. ## Outputs | Name | Bind key | Logical dtype | Rank | Shape | Description | Presence | | --- | --- | --- | --- | --- | --- | --- | | `output` | `output` | `T` | — | — | Output tensor containing the constant value(s) specified by the attribute. | required | ## Attributes Attributes and default values (overridable per request): | Attribute | Default | Description | | --- | --- | --- | | `value` | — | Tensor-valued constant represented as `{ dtype, shape, values }`; `dtype` and `shape` must match the output contract and `values` supplies the tensor elements in row-major order. | | `value_float` | — | A single float32 value producing a scalar tensor. | | `value_floats` | — | Float32 values producing a one-dimensional tensor. | ## Type constraints | Variable | Allowed dtypes | | --- | --- | | `T` | `float32`, `float16`, `int32`, `int16`, `uint32`, `int8`, `uint8`, `bool` | ## 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 - [`constant.wgsl.jinja`](build/webgpu/constant.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. ```js import { getKernel } from "@huggingface/kernels"; const kernel = await getKernel("webgpu-kernels/ai.onnx.Constant", { version: 1 }); // Explicit destinations request optional results or supply metadata that cannot be inferred. const { output } = await kernel({}, { attrs: { value_float: -2.5 }, outputs: { output: { shape: [], dtype: "float32" } }, }); ```