batch_class_ratio.md

TFSimilarity.training_metrics.batch_class_ratio

Computes the average number of examples per class within each batch.

TFSimilarity.training_metrics.batch_class_ratio(
    sampler: Sampler, num_batches: int = 100
) -> float

Similarity learning requires at least 2 examples per class in each batch. This is needed in order to construct the triplets. This function provides the average number of examples per class within each batch and can be used to check that a sampler is working correctly. The ratio should be >= 2. Args: sampler: A tf.similarity sampler object. num_batches: The number of batches to sample. Returns: The average number of examples per class.