Package org.tribuo.interop.tensorflow

Class ImageConverter

java.lang.Object
org.tribuo.interop.tensorflow.ImageConverter
All Implemented Interfaces:
com.oracle.labs.mlrg.olcut.config.Configurable, com.oracle.labs.mlrg.olcut.provenance.Provenancable<com.oracle.labs.mlrg.olcut.provenance.ConfiguredObjectProvenance>, Serializable, FeatureConverter, ProtoSerializable<org.tribuo.interop.tensorflow.protos.FeatureConverterProto>
public class ImageConverter extends Object implements FeatureConverter
Image converter. Assumes the feature id numbers are linearised ids of the form: 
 [0,0,0] = 0, [0,0,1] = 1, [0,1,0] = k, [1,0,0] = j*k, ..., [i,0,0] = i*j*k,
That is, they are in multidimensional row major order (e.g. the order used by IDXDataSource).
See Also:

  • Field Details

    • CURRENT_VERSION

      public static final int CURRENT_VERSION
      Protobuf serialization version.
      See Also:

  • Constructor Details

    • ImageConverter

      public ImageConverter(String inputName, int width, int height, int channels)
      Builds an image converter for images of the supplied size.
      Parameters:
      inputName - The input name.
      width - The image width.
      height - The image height.
      channels - The number of colour channels.

  • Method Details

    • postConfig

      public void postConfig()
      Used by the OLCUT configuration system, and should not be called by external code.
      Specified by:
      postConfig in interface com.oracle.labs.mlrg.olcut.config.Configurable

    • deserializeFromProto

      public static ImageConverter deserializeFromProto(int version, String className, com.google.protobuf.Any message) throws com.google.protobuf.InvalidProtocolBufferException
      Deserialization factory.
      Parameters:
      version - The serialized object version.
      className - The class name.
      message - The serialized data.
      Returns:
      The deserialized object.
      Throws:
      com.google.protobuf.InvalidProtocolBufferException - If the protobuf could not be parsed from the message.

    • serialize

      public org.tribuo.interop.tensorflow.protos.FeatureConverterProto serialize()
      Description copied from interface: ProtoSerializable
      Serializes this object to a protobuf.
      Specified by:
      serialize in interface ProtoSerializable<org.tribuo.interop.tensorflow.protos.FeatureConverterProto>
      Returns:
      The protobuf.

    • inputNamesSet

      public Set<String> inputNamesSet()
      Description copied from interface: FeatureConverter
      Gets a view of the names of the inputs this converter produces.
      Specified by:
      inputNamesSet in interface FeatureConverter
      Returns:
      The input names.

    • convert

      public TensorMap convert(Example<?> example, ImmutableFeatureMap featureIDMap)
      Transform implicitly pads unseen values with zero.
      Specified by:
      convert in interface FeatureConverter
      Parameters:
      example - The example to transform.
      featureIDMap - The feature id mapping to use.
      Returns:
      A 4d tensor, (1, width, height, channels) for this example.

    • convert

      public TensorMap convert(List<? extends Example<?>> examples, ImmutableFeatureMap featureIDMap)
      Transform implicitly pads unseen values with zero.

      Converts a batch of examples into a Tensor.

      Specified by:
      convert in interface FeatureConverter
      Parameters:
      examples - The examples to transform.
      featureIDMap - The feature id mapping to use.
      Returns:
      A 4d tensor, (batch-id, width, height, channels) for this example.

    • convert

      public TensorMap convert(SGDVector vector)
      Description copied from interface: FeatureConverter
      Converts a SGDVector representing the features into a TensorMap.

      It generates it as a single example minibatch.

      Specified by:
      convert in interface FeatureConverter
      Parameters:
      vector - The features to convert.
      Returns:
      A TensorMap (similar to a TF Python feed_dict) representing this vector.

    • convert

      public TensorMap convert(List<? extends SGDVector> vectors)
      Description copied from interface: FeatureConverter
      Converts a list of SGDVectors representing a batch of features into a TensorMap.
      Specified by:
      convert in interface FeatureConverter
      Parameters:
      vectors - The batch of features to convert.
      Returns:
      A TensorMap (similar to a TF Python feed_dict) representing this minibatch.

    • toString

      public String toString()
      Overrides:
      toString in class Object

    • getProvenance

      public com.oracle.labs.mlrg.olcut.provenance.ConfiguredObjectProvenance getProvenance()
      Specified by:
      getProvenance in interface com.oracle.labs.mlrg.olcut.provenance.Provenancable<com.oracle.labs.mlrg.olcut.provenance.ConfiguredObjectProvenance>