Jura Pintar
Jura Pintar
Here is [a zipped `npz` file](https://github.com/rapidsai/cugraph/files/13909474/adjacency.zip) containing the adjacency matrix of a 100,000 vertex graph. The quick-and-dirty code below will illustrate the issue: ``` import pandas as pd import cudf...
I've been testing adjacency matrices of different sizes and I'm getting some Rand index values less than 1.0 at 1100 vertices (but not at 1000 vertices). A representative run: ```...
Have you been able to make any progress on this?
Thank you for looking into this more! With version `24.04.00a49`, the smallest graph where I see randomness is 3200 nodes ([zipped `npz` file](https://github.com/rapidsai/cugraph/files/14274342/adjacency_3200.zip)). Here is a typical set of runs...
Is there any progress on this? I find myself having to run a lot of Leiden clustering these days, and it would be great to be able to reliably do...
@ChuckHastings, thank you for working on this! I've just tested with version `24.06.00a43`. The issue is resolved completely when the adjacency matrix is `float64`, but there is still some variability,...
Thank you for looking into this!
My use case is part of the one I described [in this comment on issue #48](https://github.com/Marsilea-viz/marsilea/issues/48#issuecomment-2364283153). I have scRNA-seq data from multiple experimental conditions that would best be arranged in...
This is really great! Thank you for looking into it!
Admittedly, my use case is somewhat esoteric. I've got data from a complex scRNA-seq experiment with 3 × 2 conditions, and for each condition I have a heatmap with grouped...