- Joined
- 2/14/23
- Messages
- 646
- Points
- 223
Is it a rite of passage to type up a NLA algo using numpy and then
As a follow up, is it standard to transpose all vectors upon input or creation or does everyone just accept the flipped dimensions and move on. I assume we accept and move on.
get hit with a dozen errors stimming from incorrect shape configuration in matrix multiplication
search for an hour or two wondering how they came to be and snuffing them all out
use np.transpose in many random places
finally realize that np.array() defaults to a 1xn vector but standard math notation uses nx1 vectors?
Because that's what I just did, and I'd like to think I'm not alone. As a follow up, is it standard to transpose all vectors upon input or creation or does everyone just accept the flipped dimensions and move on. I assume we accept and move on.