module `npy.xop_convert`#

Short summary#

module mlprodict.npy.xop_convert

Easier API to build onnx graphs. Inspired from skl2onnx.

Classes#

class	truncated documentation
`OnnxSubEstimator`	This operator is used to call the converter of a model to insert the node coming from the conversion into a bigger …
`OnnxSubOnnx`	This operator is used to insert existing ONNX into the ONNX graph being built.

Properties#

property	truncated documentation
`input_names`	Returns the input names.
`input_names`	Returns the input names.
`output_names`	Returns self.output_names_.
`output_names`	Returns self.output_names_.

Static Methods#

staticmethod	truncated documentation
`_to_onnx`	Converts a model into ONNX and inserts it into an ONNX graph.
`_to_onnx_sklearn`	Converts a scikit-learn model into ONNX and inserts it into an ONNX graph. The library relies on …

Methods#

method	truncated documentation
`__init__`
`__init__`
`__repr__`	usual
`__repr__`	usual
`add_to`	Adds to graph builder.
`add_to`	Adds to graph builder.
`to_onnx_this`	Returns the ONNX graph.
`to_onnx_this`	Returns the ONNX graph.

Documentation#

Easier API to build onnx graphs. Inspired from skl2onnx.

New in version 0.9.

source on GitHub

class mlprodict.npy.xop_convert.OnnxSubEstimator(model, *inputs, op_version=None, output_names=None, options=None, initial_types=None, **kwargs)#

Bases: OnnxSubOnnx

This operator is used to call the converter of a model to insert the node coming from the conversion into a bigger ONNX graph. It supports model from scikit-learn using sklearn-onnx.

Parameters:

model – model to convert
inputs – inputs
op_version – targetted opset
options – to rewrite the options used to convert the model
initial_types – the implementation may be wrong in guessing the input types of the model, this parameter can be used to overwrite them, usually a dictionary { input_name: numpy array as an example }
kwargs – any other parameters such as black listed or white listed operators

source on GitHub

__init__(model, *inputs, op_version=None, output_names=None, options=None, initial_types=None, **kwargs)#

__repr__()#: usual

static _to_onnx(model, inputs, op_version=None, options=None, initial_types=None, **kwargs)#

Converts a model into ONNX and inserts it into an ONNX graph.

Parameters:

model – a trained machine learned model
inputs – inputs
op_version – opset versions or None to use the latest one
options – options to change the behaviour of the converter
kwargs – additional parameters such as black listed or while listed operators

Returns:

ONNX model

The method currently supports models trained with scikit-learn, xgboost, :epkg`:lightgbm`.

source on GitHub

static _to_onnx_sklearn(model, inputs, op_version=None, options=None, initial_types=None, **kwargs)#

Converts a scikit-learn model into ONNX and inserts it into an ONNX graph. The library relies on function to_onnx and library :epkg:`skearn-onnx`.

Parameters:

model – a trained machine learned model
inputs – inputs
op_version – opset versions or None to use the latest one
initial_types – if None, the input types are guessed from the inputs. The function converts into ONNX the previous node of the graph and tries to infer the initial_types with the little informations it has. It may not work. It is recommended to specify this parameter.
options – options to change the behaviour of the converter
kwargs – additional parameters such as black listed or while listed operators

Returns: