Crate cubecl_random

Modules§

assert_at_least_one_value_per_bin: Asserts that there is at least one value per bin
assert_mean_approx_equal: Asserts that the mean of a dataset is approximately equal to an expected value, within 2.5 standard deviations. There is a very small chance this raises a false negative.
assert_normal_respects_68_95_99_rule: Asserts that the distribution follows the 68-95-99 rule of normal distributions, following the given mean and standard deviation.
assert_number_of_1_proportional_to_prob: For a bernoulli distribution: asserts that the proportion of 1s to 0s is approximately equal to the expected probability.
assert_wald_wolfowitz_runs_test: Asserts that the elements of the data, sorted into two bins, are elements of the sequence are mutually independent. There is a very small chance it gives a false negative.
calculate_bin_stats: Sorts the data into bins for ranges of equal sizes
random_bernoulli: Pseudo-random generator with bernoulli distribution
random_normal: Pseudo-random generator with uniform distribution
random_uniform: Pseudo-random generator with uniform distribution
seed
to_unit_interval_closed_open: Converts a u32 into a f32 in the unit interval [0.0, 1.0). Used for generating random floats.
to_unit_interval_open: Converts a u32 into a f32 in the unit interval (0.0, 1.0). Used for generating random floats.