Oleksandr Pavlyk
Oleksandr Pavlyk
I would love to.
@PeterDykas The support has been added in [nanobind](https://github.com/wjakob/nanobind). This issue is to add support for DLPack casters for pybind11 following suite.
You may find the article of Tamara Kolda "[Tensor Decompositions and Applications](https://epubs.siam.org/doi/pdf/10.1137/07070111X)", from SIAM Review, Vol. 51, No. 3, pp. 455–500 of interest.
Compressed sparse fiber is another recent format to consider: http://glaros.dtc.umn.edu/gkhome/node/1177
> @leofang While there is no stateless PRNG, HW-based true random number generators have empty state (MKL's NONDETERM basic random number generator) and these should be supported by the spec...
> I can go into that in more detail here or somewhere else, if anyone would like. I vote for presenting them here, please.
My issue with permitting broadcasting of 0d and 1d arrays which contain 1 element to empty arrays that have no elements is with the loss of data: ``` In [1]:...
@kgryte Empty arrays are not consistently 'poisonous`. This behavior is specific when other array has size 1: ``` In [1]: import numpy as np In [2]: np.array([1]) + np.array([]) Out[2]:...
:+1: for the adoption of symmetric proximity comparator. > All library implementations add a `equal_nan: bool = False` flag to the functions. Can you indicate where this might be useful?...
Scikit-Learn developers are taking a look at NEP-37 vs. NEP-18: https://github.com/scikit-learn/scikit-learn/pull/17676