pyprobound.utils

Typed helper functions.

Functions

avg_pool1d(tensor[, kernel])	Average pooling along the first dimension.
betaln(z_1, z_2)	Computes the natural logarithm of the beta function.
ceil_div(dividend, divisor)	Typed ceiling division.
clear_cache()	Calls the empty_cache() function matching the available GPU backends.
count_kmers(sequences[, kmer_length, vocabulary])	Returns a sparse count matrix of k-mers in a list of sequences.
floor_div(dividend, divisor)	Typed floor division.
get_ordinal(integer)	Converts an integer to a string with an ordinal suffix.
get_split_size(max_embedding_size, ...)	Calculates the minibatch needed to avoid GPU limits on tensor sizes.
log1mexp(tensor, /[, eps])	Computes the element-wise log1mexp in a numerically stable way.