Naim

I will take a look at it tomorrow

@jpintar Could you please provide us with the smallest adjacency matrix [perhaps sub-sampled] where cluster assignments is different, ie adjusted Rand index is less than 1.0 ?

For the adjacency matrix that you provided, locally I get similar modularity (differs by 1e-03 from run to run)

@jpintar @mbruhns Thank you for reporting the issue. We have a fix that got merged in `brach-24.04`, and one can use `brach-24.04` nightly to run the latest version of the...

Thank you for checking out the latest version. We're still investigating it to find out the remaining cause of the variation. We'll get back to you as soon as we...

Hi, We'll review the code again, aiming to eliminate any additional sources of randomness other than numerical precision. I hope to get back to you soon.

/okay to test

/okay to test

/okay to test

/okay to test