17

I think I've hit an all-time record of being a superior and condescending engineer, when I casually remarked that we are wrong to use L1 loss function on the yaw, because you can't take Euclidean-like distance on a cylinder.

Comments
Add Comment