←back to thread

I don't like NumPy

(dynomight.net)
480 points MinimalAction | 1 comments | | HN request time: 0.209s | source
1. QuadmasterXLII ◴[] No.44000256[source]
Numpy and torch have opposite conventions as to which end to broadcast from if arrays hve different numbers of dimensions, and as a result I completely forgot both. At this point I have to always None both arrays to the same dimension, which is ugly but actually fairly explicit