Leland McInnes
Leland McInnes
I would be happy to see a pull request to implement such a thing, but given that a user can work around these issues on their end I don't see...
The most likely cause would be an all zeros vector. Problems are infeasible if there is no way to transport one probability distribution to the other. If one of the...
This seems to be related to how Numba handles things. Hopefully there is a reasonable work-around. It will take some time to figure out the right way to handle this.
Potentially I have a fix here. Hopefully this works... I won't have time to test it for a while as I'm travelling very soon.
Thanks. I believe the current master has fixes in for this from @jamestwebber so hopefully this will be resolved when I get a new patch release out.
I think there is a version mistmatch between what saved the model and what is loading the model. There were some changes in recent versions to allow for, for example,...
That should work; I'll see if I can look into it.
Due to how things have to get handled by numba, in practice you need to pass *all* the required arguments in the correct order to the distance function. That means...
There is a paper on that, and things related (arXiv:1908.00814v3); currently that is in the list of things for me to do, but I'm not sure when or if that...
I would be happy to see any efforts you want to make on this. For your specific problem you could, in principle, build the index on each dataset and then...