Edward Middleton

Thank you for doing this @jturner314 this is awesome. I really like the optimization with the break after the first singular value below the threshold. Is this something we could...

I am working through this, I noticed octave[1] calculates the threshold using ε⋅max(m, n)⋅max(Σ) but NumPy[2] just does rcond⋅max(Σ) where rcond is an argument that defaults to 1e-15, which is...

I was thinking of having an optional argument rcond like NumPy that defaults to ε⋅max(m, n) if its not specified but I would really like to know where the 1e-15...

The plot thickens, numpy.linalg.matrix_rank uses the threshold ε⋅max(m, n)⋅max(Σ) and references W. H. Press, S. A. Teukolsky, W. T. Vetterling and B. P. Flannery, "Numerical Recipes (3rd edition)", Cambridge University...

I have implemented tests for various ranks for matrix but it seems like there must be a better way to create them.

Thanks, Is there any way to run locally with the unreleased version of ndarray-linalg. That was what I was trying to do as I am using some features in ndarray...

Just as a follow up on this, I was not able to get this working without explicitly downloading the source from github and including it by path. i.e. ```toml [dependencies.ndarray]...

@cedeerwe I could only get quickcheck to work with ```toml quickcheck = "0.9" quickcheck_macros = "0.9" ```

The functionality to use this has not yet been implemented.

The server seems to be down at the moment.