Open marianotepper opened 8 years ago
In utils.linalg.norm, why not use the already implemented norm functions in scipy.sparse.linalg, instead of re-implementing everything from scratch? This would allow to seamlessly add the axis parameter, which is very handy oftentimes.
In utils.linalg.norm, why not use the already implemented norm functions in scipy.sparse.linalg, instead of re-implementing everything from scratch? This would allow to seamlessly add the axis parameter, which is very handy oftentimes.