Module: tfr.utils

Utility functions for ranking library.

Functions

de_noise(...): Returns a float Tensor as the de-noised counts.

gather_per_row(...): Gathers the values from input tensor based on per-row indices.

is_label_valid(...): Returns a boolean Tensor for label validity.

organize_valid_indices(...): Organizes indices in such a way that valid items appear first.

padded_nd_indices(...): Pads the invalid entries by valid ones and returns the nd_indices.

parse_keys_and_weights(...): Parses the encoded key to keys and weights.

ragged_to_dense(...): Converts given inputs from ragged tensors to dense tensors.

reshape_first_ndims(...): Reshapes the first n dims of the input tensor to new shape.

reshape_to_2d(...): Converts the given tensor to a 2-D Tensor.

shuffle_valid_indices(...): Returns a shuffle of indices with valid ones on top.

sort_by_scores(...): Sorts list of features according to per-example scores.

sorted_ranks(...): Returns an int Tensor as the ranks (1-based) after sorting scores.

Type Aliases

LossFunction

MetricFunction

TensorLike

TransformationFunction