Module: TFSimilarity.callbacks

Specialized callbacks that track similarity metrics during training

Classes

class Callback: Abstract base class used to build new callbacks.
class ClassificationMatch: Abstract base class for defining the classification matching strategy.
class ClassificationMetric: Abstract base class for computing classification metrics.
class EvalCallback: Epoch end evaluation callback that build a test index and evaluate
class Evaluator: Evaluates search index performance and calibrates it.
class FloatTensor: Float tensor
class IntTensor: Integer tensor
class MemoryEvaluator: In memory index performance evaluation and classification.
class Path: PurePath subclass that can make system calls.
class SimilarityModel: Specialized Keras.Model which implement the core features needed for
class Tensor: A tf.Tensor represents a multidimensional array of elements.

SplitValidationLoss(...): Creates the validation callbacks.
make_classification_metric(...): Convert classification metric from str name to object if needed.
unpack_lookup_distances(...)
unpack_lookup_labels(...)
unpack_results(...): Updates logs, writes summary, and returns list of strings of