ValueError(f'negative axis {i} index: {idx.min()}')

1

Hello everyone!
I’m trying to solve a convex optimization problem with a inequality constraint on cvx. The objective is just a frobenious norm of matrices produtc in a mse fashion. I have an error like:
raise ValueError(f’negative axis {i} index: {idx.min()}')
ValueError: negative axis 0 index: -2147458290
when passing the whole X matrix, that is a huge matrix of 45000 columns. I guess that this error is due to the large scale of X matrix, this do not happen if I pass just 100 columns, for example, of X. Suggestions to solve this problem?

2

This is a forum for CVX in Matlab.

Your question is about CVXPY in Python and is addressed here: ValueError: negative axis 0 index for large dimensional variables · Issue #2698 · cvxpy/cvxpy · GitHub

1 Like