{
  "$schema": "https://ambiqai.github.io/helia-ui/schema/reference-model-1.json",
  "generatedFrom": {
    "tool": "pyref",
    "version": "1.7.3"
  },
  "language": "python",
  "modules": [
    {
      "description": "Checking a run's outputs: reference outputs from host LiteRT or caller files, compared tensor by tensor, with a verdict only when a tolerance is given.\n\nEvery name on this page is imported from `helia_profiler`.\n\n**API tier:** `experimental`\n\nGenerated from the `src/helia_profiler` tree `f38e378c801d34438affb680257217e0b58d2cd1`.",
      "name": "expect",
      "path": "helia_profiler.expect",
      "submodules": [],
      "summary": "Checking a run's outputs: reference outputs from host LiteRT or caller files, compared tensor by tensor, with a verdict only when a tolerance is given.",
      "symbols": [
        {
          "description": "`dataclass`\n\nLargest accepted absolute error per element, in each tensor's stored units.\n\nInteger tensors count quantization steps; float tensors count values. One\ntolerance applies to every output tensor.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.Tolerance",
          "kind": "class",
          "language": "python",
          "members": [
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Tolerance.max_abs_error",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "max_abs_error",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "max_abs_error: float = 0.0",
              "source": {
                "line": 53,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L53"
              },
              "summary": ""
            }
          ],
          "name": "Tolerance",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "Tolerance(max_abs_error: float = 0.0) -> None",
          "source": {
            "line": 45,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L45"
          },
          "summary": "Largest accepted absolute error per element, in each tensor's stored units."
        },
        {
          "description": "`dataclass`\n\nWhere reference outputs came from.\n\nA host reference records the LiteRT package version, its op resolver and\nhow many times the model was invoked on the inputs.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.ReferenceOrigin",
          "kind": "class",
          "language": "python",
          "members": [
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.ReferenceOrigin.kind",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "kind",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "kind: ReferenceKind",
              "source": {
                "line": 68,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L68"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.ReferenceOrigin.litert_version",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "litert_version",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "litert_version: str | None = None",
              "source": {
                "line": 69,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L69"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.ReferenceOrigin.resolver",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "resolver",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "resolver: str | None = None",
              "source": {
                "line": 70,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L70"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.ReferenceOrigin.invocations",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "invocations",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "invocations: int | None = None",
              "source": {
                "line": 71,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L71"
              },
              "summary": ""
            }
          ],
          "name": "ReferenceOrigin",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "ReferenceOrigin(\n    kind: ReferenceKind,\n    litert_version: str | None = None,\n    resolver: str | None = None,\n    invocations: int | None = None,\n) -> None",
          "source": {
            "line": 60,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L60"
          },
          "summary": "Where reference outputs came from."
        },
        {
          "description": "`dataclass`\n\nReference bytes for every output, in model order, and their origin.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.ReferenceOutputs",
          "kind": "class",
          "language": "python",
          "members": [
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.ReferenceOutputs.origin",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "origin",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "origin: ReferenceOrigin",
              "source": {
                "line": 78,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L78"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.ReferenceOutputs.outputs",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "outputs",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "outputs: tuple[bytes, ...]",
              "source": {
                "line": 79,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L79"
              },
              "summary": ""
            }
          ],
          "name": "ReferenceOutputs",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "ReferenceOutputs(origin: ReferenceOrigin, outputs: tuple[bytes, ...]) -> None",
          "source": {
            "line": 74,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L74"
          },
          "summary": "Reference bytes for every output, in model order, and their origin."
        },
        {
          "description": "`dataclass`\n\nWhat a run's outputs are compared against.\n\nBuild with :meth:`host_litert`, :meth:`from_build` or :meth:`from_files`.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.Expect",
          "kind": "class",
          "language": "python",
          "members": [
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Expect.source",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "source",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "source: ReferenceKind",
              "source": {
                "line": 89,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L89"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Expect.files",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "files",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "files: tuple[FixtureFile | Path, ...] = ()",
              "source": {
                "line": 90,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L90"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Expect.resolver",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "resolver",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "resolver: LiteRtResolver = 'BUILTIN_REF'",
              "source": {
                "line": 91,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L91"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Expect.tolerance",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "tolerance",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "tolerance: Tolerance | None = None",
              "source": {
                "line": 92,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L92"
              },
              "summary": ""
            },
            {
              "description": "`classmethod`\n\nRun the model with host LiteRT on the run's input bytes (LiteRT models only).",
              "examples": [],
              "id": "helia_profiler.Expect.host_litert",
              "kind": "method",
              "language": "python",
              "members": [],
              "name": "host_litert",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "host_litert(*, resolver: LiteRtResolver = 'BUILTIN_REF', tolerance: Tolerance | None = None) -> Expect",
              "source": {
                "line": 94,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L94"
              },
              "summary": "Run the model with host LiteRT on the run's input bytes (LiteRT models only)."
            },
            {
              "description": "`classmethod`\n\nUse the expected outputs the fixture build carries.",
              "examples": [],
              "id": "helia_profiler.Expect.from_build",
              "kind": "method",
              "language": "python",
              "members": [],
              "name": "from_build",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "from_build(*, tolerance: Tolerance | None = None) -> Expect",
              "source": {
                "line": 102,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L102"
              },
              "summary": "Use the expected outputs the fixture build carries."
            },
            {
              "description": "`classmethod`\n\nCompare with caller bytes, one file per output in model order.",
              "examples": [],
              "id": "helia_profiler.Expect.from_files",
              "kind": "method",
              "language": "python",
              "members": [],
              "name": "from_files",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "from_files(*files: FixtureFile | str | Path = (), tolerance: Tolerance | None = None) -> Expect",
              "source": {
                "line": 107,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L107"
              },
              "summary": "Compare with caller bytes, one file per output in model order."
            }
          ],
          "name": "Expect",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "Expect(\n    source: ReferenceKind,\n    files: tuple[FixtureFile | Path, ...] = (),\n    resolver: LiteRtResolver = 'BUILTIN_REF',\n    tolerance: Tolerance | None = None,\n) -> None",
          "source": {
            "line": 82,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L82"
          },
          "summary": "What a run's outputs are compared against."
        },
        {
          "description": "`dataclass`\n\nComparison facts for one output tensor.\n\n``max_abs_error`` is in stored units; ``max_abs_error_real`` is the same\nerror after dequantization, ``None`` for an unquantized tensor. A NaN on\none side only counts as a mismatch with infinite error. ``within_tolerance``\nis ``None`` when no tolerance was given.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.TensorComparison",
          "kind": "class",
          "language": "python",
          "members": [
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.name",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "name",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "name: str",
              "source": {
                "line": 128,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L128"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.index",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "index",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "index: int",
              "source": {
                "line": 129,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L129"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.dtype",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "dtype",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "dtype: str",
              "source": {
                "line": 130,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L130"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.shape",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "shape",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "shape: tuple[int, ...]",
              "source": {
                "line": 131,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L131"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.reference_sha256",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "reference_sha256",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "reference_sha256: str",
              "source": {
                "line": 132,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L132"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.observed_sha256",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "observed_sha256",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "observed_sha256: str",
              "source": {
                "line": 133,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L133"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.elements",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "elements",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "elements: int",
              "source": {
                "line": 134,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L134"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.mismatched_elements",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "mismatched_elements",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "mismatched_elements: int",
              "source": {
                "line": 135,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L135"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.max_abs_error",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "max_abs_error",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "max_abs_error: float",
              "source": {
                "line": 136,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L136"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.max_abs_error_real",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "max_abs_error_real",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "max_abs_error_real: float | None",
              "source": {
                "line": 137,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L137"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.within_tolerance",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "within_tolerance",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "within_tolerance: bool | None",
              "source": {
                "line": 138,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L138"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.TensorComparison.identical",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "identical",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "identical: bool",
              "source": {
                "line": 141,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L141"
              },
              "summary": ""
            }
          ],
          "name": "TensorComparison",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "TensorComparison(\n    name: str,\n    index: int,\n    dtype: str,\n    shape: tuple[int, ...],\n    reference_sha256: str,\n    observed_sha256: str,\n    elements: int,\n    mismatched_elements: int,\n    max_abs_error: float,\n    max_abs_error_real: float | None,\n    within_tolerance: bool | None,\n) -> None",
          "source": {
            "line": 118,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L118"
          },
          "summary": "Comparison facts for one output tensor."
        },
        {
          "description": "`dataclass`\n\nPer-tensor comparison facts, the reference they used and an optional verdict.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.Verification",
          "kind": "class",
          "language": "python",
          "members": [
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Verification.reference",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "reference",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "reference: ReferenceOutputs",
              "source": {
                "line": 149,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L149"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Verification.tensors",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "tensors",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "tensors: tuple[TensorComparison, ...]",
              "source": {
                "line": 150,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L150"
              },
              "summary": ""
            },
            {
              "description": "",
              "examples": [],
              "id": "helia_profiler.Verification.tolerance",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "tolerance",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "tolerance: Tolerance | None",
              "source": {
                "line": 151,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L151"
              },
              "summary": ""
            },
            {
              "description": "Whether every tensor is within tolerance; ``None`` when no tolerance was given.",
              "examples": [],
              "id": "helia_profiler.Verification.matched",
              "kind": "attribute",
              "language": "python",
              "members": [],
              "name": "matched",
              "params": [],
              "raises": [],
              "returns": [],
              "signature": "matched: bool | None",
              "source": {
                "line": 154,
                "path": "src/helia_profiler/runs/expect.py",
                "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L154"
              },
              "summary": "Whether every tensor is within tolerance; None when no tolerance was given."
            }
          ],
          "name": "Verification",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "Verification(reference: ReferenceOutputs, tensors: tuple[TensorComparison, ...], tolerance: Tolerance | None) -> None",
          "source": {
            "line": 145,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L145"
          },
          "summary": "Per-tensor comparison facts, the reference they used and an optional verdict."
        },
        {
          "description": "Run a LiteRT model on the host and return its output bytes in *outputs* order.\n\n*inputs* must cover every model input. They are restored before every\ninvocation, as the fixture firmware does, so internal state carries across\n*invocations* while inputs do not.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.host_reference",
          "kind": "function",
          "language": "python",
          "members": [],
          "name": "host_reference",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "host_reference(\n    model: bytes,\n    inputs: Sequence[tuple[FixtureTensor, bytes]],\n    outputs: Sequence[FixtureTensor],\n    *,\n    resolver: LiteRtResolver = 'BUILTIN_REF',\n    invocations: int = 1,\n) -> ReferenceOutputs",
          "source": {
            "line": 161,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L161"
          },
          "summary": "Run a LiteRT model on the host and return its output bytes in outputs order."
        },
        {
          "description": "Compare observed output bytes with reference outputs, tensor by tensor.\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.compare_outputs",
          "kind": "function",
          "language": "python",
          "members": [],
          "name": "compare_outputs",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "compare_outputs(\n    tensors: Sequence[FixtureTensor],\n    reference: ReferenceOutputs,\n    observed: Sequence[bytes],\n    tolerance: Tolerance | None = None,\n) -> Verification",
          "source": {
            "line": 235,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L235"
          },
          "summary": "Compare observed output bytes with reference outputs, tensor by tensor."
        },
        {
          "description": "Compare a fixture capture's outputs with *expect*.\n\nA host reference runs the build's model on its pinned input bytes as many\ntimes as the firmware invokes it (warmups, then timed iterations).\n\n**API tier:** `experimental`",
          "examples": [],
          "id": "helia_profiler.compare_fixture_outputs",
          "kind": "function",
          "language": "python",
          "members": [],
          "name": "compare_fixture_outputs",
          "params": [],
          "raises": [],
          "returns": [],
          "signature": "compare_fixture_outputs(build: FixtureBuild, capture: FixtureCaptureResult, expect: Expect) -> Verification",
          "source": {
            "line": 304,
            "path": "src/helia_profiler/runs/expect.py",
            "url": "https://github.com/AmbiqAI/helia-profiler/blob/main/src/helia_profiler/runs/expect.py#L304"
          },
          "summary": "Compare a fixture capture's outputs with expect."
        }
      ]
    }
  ],
  "name": "helia_profiler"
}