Daniel Coquelin

the recommended way to get `black` running in a way which is unified with the codebase is to use `pre-commit` (`pip install pre-commit` or ` conda install -c conda-forge pre-commit...

what is your data size for the benchmarks?

okay i was just curious. the `**` is very bad. but this is a direct call to the binary op. im not sure why it would take so much longer...

or maybe we just dont use binary op in pow

I cant speak for Claudia but I did a bit of testing and with a minor rewrite of the pow function i was able to close the gap a bit:...

The OpenMP issue is interesting. We do not do anything with threading and rely on torch to handle this for us. During some internal testing, we found that the optimum...

rerun tests

Im not seeing the changes that you made when you moved the create partition interface function over. as for the `get` field, you mean something which gets the data of...

ive implemented that now. i tested it with a small example: ```python x = ht.arange(8* 3* 2).reshape((8, 3, 2)).resplit(0) print(x.__partitioned__['get']((0, 0, 0))) ``` ``` [1] None [2] None [0] tensor([[[...

@fschlimb I have added a bit more functionality to `from_partitioned`. It now supports non-zero split axes and i have also added a `from_partition_dict` function which does the same thing, but...