.. blogpost::
    :title: ONNX, runtime
    :keywords: onnx, onnxrt
    :date: 2019-06-25
    :categories: onnx

    Somebody asked me one day if it would be difficult to
    write a runtime for :epkg:`ONNX` in :epkg:`Rust`.
    I just replied that it should not take that long
    but it would require to implement a way to goes
    through the nodes of the :epkg:`ONNX` graph
    and to have an implementation for every
    :epkg:`ONNX Operators`...

    So tried to do it in :epkg:`Python` only
    for :epkg:`scikit-learn` models. Well...
    It is not that difficult. And yet was class
    :class:`OnnxInference
    <mlprodict.onnxrt.onnx_inference.OnnxInference>`
    with one method :meth:`to_sequence
    <mlprodict.onnxrt.onnx_inference.OnnxInference.to_sequence>`.
    Implementation of every operator needed is in folder
    `onnxrt <https://github.com/sdpython/mlprodict/tree/master/mlprodict/onnxrt>`_.

    Page about :ref:`l-CMD2` might be interesting to
    compare runtime.