ONNXRuntimeAdapter

class ONNXRuntimeAdapter

Bases: object

Wrapper around ONNX-runtime that behaves like a PyTorch nn.Module.

That is, it implements __call__() and accepts PyTorch Tensors as inputs and also outputs PyTorch Tensors.

__init__(ort_session: ort.InferenceSession) → None

Constructor.

Parameters:

ort_session (ort.InferenceSession) – ONNX-runtime InferenceSession.

Return type:

None

Methods

__init__(ort_session)	Constructor.
from_file(path[, providers])	Construct from file.

__init__(ort_session: ort.InferenceSession) → None

Constructor.

Parameters:

ort_session (ort.InferenceSession) – ONNX-runtime InferenceSession.

Return type:

None

classmethod from_file(path: str, providers: list[str] | None = None) → Self

Construct from file.

Parameters:

• path (str) – Path to a .onnx file.

• providers (list[str] | None) – ONNX-runtime execution providers. See onnxruntime documentation for more details. Defaults to None.

Returns:

An ONNXRuntimeAdapter instance.

Return type:

ONNXRuntimeAdapter