tensorfx.models

Machine Learning Scenarios

class tensorfx.models.ClassificationScenario(labels)

Represents a classification machine learning scenario.

indices_to_labels(indices, name='label')

Converts the specified integer indices into equivalent string label names.

The implementation builds a HashTable, to perform a lookup operation.

Parameters:indices – the int64 indices to convert into equivalent strings.
Returns:A set of tensors representing the labels by name.
labels

Retrieves the labels being predicted.

labels_to_indices(strings, one_hot, name='indices')

Converts the specified string label names into equivalent indices.

The implementation builds a HashTable, to perform a lookup operation.

Parameters:
  • strings – the string labels to convert into equivalent indices.
  • one_hot – whether to convert the indices into their one-hot representation.
Returns:

A set of tensors representing the labels by indices.

num_labels

Retrieves the number of labels being predicted.