search

aneeshnaik / lintsampler

Efficient random sampling via linear interpolation.
https://lintsampler.readthedocs.io
MIT License
1 stars 1 forks source link

testing DensityTree; try upgrades #4

Closed michael-petersen closed 1 month ago

michael-petersen commented 1 month ago

I did a quick addition of the other pdf parameters (vectorizedpdf, pdf_args,pdf_kwargs). I also tried a quick implementation in the DM notebook; however, I'm running into probable user error, where the refinement takes a very long time.

My guess is that I don't understand the parameters quite right; perhaps you could take a quick look at the notebook? You can also cherrypick anything you like from the tree.py upgrades; I think we'll want that parallel structure in the longer term.

aneeshnaik commented 1 month ago

The changes to tree.py all look great to me --- I'll merge this PR with all the changes. RE: the DM notebook, I'll have a go at running it after I've merged the PR. Might well just be that my implementation of the tree refinement is a little slow.