Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

multivariate_normal sample fails for large matrix

See original GitHub issue

from numpy.random import RandomState

s = RandomState(0)

N = 50000
m = s.randn(N)
G = s.randn(N, 100)
K = G.dot(G.T)

u = s.multivariate_normal(m, K)

prints init_dgesdd failed init.

Additional information

> python --version
Python 3.5.2 :: Anaconda 4.2.0 (64-bit)
> python -c "import numpy; print(numpy.__version__)"
1.11.1
> python -c "import scipy; print(scipy.__version__)"
0.18.1

Issue Analytics

State:
Created 7 years ago
Comments:6 (3 by maintainers)

Top GitHub Comments

1reaction

pvcommented, Nov 9, 2016

Numpy also does not currently support lapack64: gh-5906

0reactions

pvcommented, Nov 9, 2016

LAPACK is typically compiled with 32-bit integers. I don’t know if any linux distribution offers 64-bit lapack as a package (e.g. ubuntu does not). . But Numpy should raise exception for failed LAPACK init, not just print error messages.

Top Results From Across the Web

Lesson 4: Multivariate Normal Distribution - STAT ONLINE

This lesson is concerned with the multivariate normal distribution. ... is going to be approximately multivariate normally distributed for large samples.

Sampling from a Multivariate Normal Distribution

In this post I want to describe how to sample from a multivariate normal distribution following section A.2 Gaussian Identities of the book...

mvtnorm: Multivariate Normal and t Distributions

If the degrees of freedom ν (the argument df) is larger than 2, then Cov(X) = Σν/(ν − 2). Furthermore, in this case...

Why Multivariate Distribution with Covariance Matrix with ...

The recipe you're using assumes that your covariance matrix is full rank. Cholesky decomposition is only unique if Σ is positive definite.

MemoryError when generating numpy.MultivariateNormal

How could I generate a random dataset this large with these specifications without getting this error? Edit: Here's my stack trace: Traceback (...