Introducing LiteRT: Google's high-performance runtime for on-device AI, formerly known as TensorFlow Lite. Learn more

EmbeddingResult

public abstract class EmbeddingResult

Represents the embedding results of a model. Typically used as a result for embedding tasks.

Public Constructors

EmbeddingResult()

Public Methods

static EmbeddingResult	create(List<Embedding> embeddings, Optional<Long> timestampMs) Creates a `EmbeddingResult` instance.
static EmbeddingResult	createFromProto(EmbeddingsProto.EmbeddingResult proto) Creates a `EmbeddingResult` object from a `ERROR(/EmbeddingsProto.EmbeddingResult)` protobuf message.
abstract List<Embedding>	embeddings() The embedding results for each head of the model.
abstract Optional<Long>	timestampMs() The optional timestamp (in milliseconds) of the start of the chunk of data corresponding to these results.

Inherited Methods

From class java.lang.Object

boolean	equals(Object arg0)
final Class<?>	getClass()
int	hashCode()
final void	notify()
final void	notifyAll()
String	toString()
final void	wait(long arg0, int arg1)
final void	wait(long arg0)
final void	wait()

Public Constructors

public EmbeddingResult ()

Public Methods

public static EmbeddingResult create (List<Embedding> embeddings, Optional<Long> timestampMs)

Creates a EmbeddingResult instance.

Parameters

embeddings	the list of `Embedding` objects containing the embedding for each head of the model.
timestampMs	the optional timestamp (in milliseconds) of the start of the chunk of data corresponding to these results.

public static EmbeddingResult createFromProto (EmbeddingsProto.EmbeddingResult proto)

Creates a EmbeddingResult object from a ERROR(/EmbeddingsProto.EmbeddingResult) protobuf message.

Parameters

proto	the `ERROR(/EmbeddingsProto.EmbeddingResult)` protobuf message to convert.

public abstract List<Embedding> embeddings ()

The embedding results for each head of the model.

public abstract Optional<Long> timestampMs ()

The optional timestamp (in milliseconds) of the start of the chunk of data corresponding to these results.

This is only used for embedding extraction on time series (e.g. audio embedder). In these use cases, the amount of data to process might exceed the maximum size that the model can process: to solve this, the input data is split into multiple chunks starting at different timestamps.