Sterling G. Baird

@BastianZim this is great! Thanks for letting me know, and good to know about GitHub search. Didn't realize that - will probably stick with VS Code Ctrl+Shift+F when I want...

The online demo for chemdataextractor2 seems to work fine. http://www.chemdataextractor2.org/results/b872ca31-7e62-49de-b1b9-41dcc969b3f0

> Maybe I could pass a tuple of (`fixture`, `attr`, `check_value`) with `indirect=False`. Assuming that works, will I be losing the benefit of using fixtures in the first place? Not...

Any Windows users that have run this locally yet?

https://github.com/lmcinnes/pynndescent/issues/136#issuecomment-994289807 I think [dist_matrix](https://github.com/sparks-baird/dist-matrix) can be useful in this context, but only for 1D feature vectors, i.e. it's not general to nD like the optimal transport libraries., but it's very...

I'm noticing that very small datasets (e.g. 100 points) seems to be slower than slightly larger datasets (e.g. 800 points) according to the [docs: benchmarking results](https://umap-learn.readthedocs.io/en/latest/benchmarking.html), though I haven't been...

Perhaps I should use PCA as a fast swap-in, and then switch back to UMAP for production purposes. For densMAP values (i.e. radius), I might just need to use a...

Thanks for checking into that. I'll try to get some MWEs and see what happens.

https://stackoverflow.com/a/62550189/13697228 mentions data length needing to be divisible by `batch_size`. Changed `batch_size` to 1 everywhere and same issue. Here's the log: ```python 22-06-09 23:28:39.190 - INFO: Create the log file...

@Janspiry if you close the issue, the person that originally opened it can't reopen the issue. How do you suggest I fix the error, `Caught IndexError in DataLoader worker process...