Microsoft.ML.Transforms.Onnx Namespace

Namespace containing ONNX model loading and transformation components.

Classes

DnnImageFeaturizerEstimator

Applies a pre-trained deep neural network (DNN) model to featurize input image data.

DnnImageFeaturizerInput

Helper class for storing all the inputs to an extension method on a DnnImageModelSelector required to return a chain of two OnnxScoringEstimator.

DnnImageModelSelector

Helper class for selecting a pre-trained DNN image featurization model to use in the DnnImageFeaturizerEstimator.

OnnxMapType

The corresponding Type of ONNX's map type in IDataView's type system. In other words, if an ONNX model produces a map, a column in IDataView may be typed to OnnxMapType. Its underlying type is IDictionary<TKey,TValue>, where the generic type "TKey" and "TValue" are the input arguments of OnnxMapType(Type, Type).

OnnxMapTypeAttribute

To declare OnnxMapType column in IDataView as a field in a class, the associated field should be marked with OnnxMapTypeAttribute. Its uses are similar to those of VectorTypeAttribute and other classes derived from DataViewTypeAttribute.

OnnxScoringEstimator

IEstimator<TTransformer> for scoring ONNX models in the ML.NET framework.

OnnxSequenceType

The corresponding Type of ONNX's sequence type in IDataView's type system. In other words, if an ONNX model produces a sequence, a column in IDataView may be typed to OnnxSequenceType. Its underlying type is IEnumerable<T>, where the generic type "T" is the input argument of OnnxSequenceType(Type).

OnnxSequenceTypeAttribute

To declare OnnxSequenceType column in IDataView as a field in a class, the associated field should be marked with OnnxSequenceTypeAttribute. Its uses are similar to those of VectorTypeAttribute and other classes derived from DataViewTypeAttribute.

OnnxTransformer

ITransformer resulting from fitting an OnnxScoringEstimator. Please refer to OnnxScoringEstimator to learn more about the necessary dependencies, and how to run it on a GPU.