ResNet18Extension Class

Definition

Namespace:

Microsoft.ML

Assembly:

Microsoft.ML.DnnImageFeaturizer.ResNet18.dll

Package:

Microsoft.ML.DnnImageFeaturizer.ResNet18 v0.21.1

This is an extension method to be used with the DnnImageFeaturizerEstimator in order to use a pretrained ResNet18 model. The NuGet containing this extension is also guaranteed to include the binary model file.

public static class ResNet18Extension

type ResNet18Extension = class

Public Module ResNet18Extension

Inheritance

Object

ResNet18Extension

Methods

ResNet18(DnnImageModelSelector, IHostEnvironment, String, String)

Returns an estimator chain with the two corresponding models (a preprocessing one and a main one) required for the ResNet pipeline. Also includes the renaming ColumnsCopyingTransforms required to be able to use arbitrary input and output column names. This assumes both of the models are in the same location as the file containing this method, which they will be if used through the NuGet. This should be the default way to use ResNet18 if importing the model from a NuGet.

ResNet18(DnnImageModelSelector, IHostEnvironment, String, String, String)

This allows a custom model location to be specified. This is useful is a custom model is specified, or if the model is desired to be placed or shipped separately in a different folder from the main application. Note that because ONNX models must be in a directory all by themselves for the OnnxTransformer to work, this method appends a ResNet18Onnx/ResNetPrepOnnx subdirectory to the passed in directory to prevent having to make that directory manually each time.