假设
a = np.random.rand(np.random.randint(10), 4)
b = np.random.rand(np.random.randint(10), 4)
我想向量化以下内容：
from itertools import product
res = np.array([foo(a[i, :], b) for i in range(a.shape[0])])
foo can take a vector and a matrix and yield a correct result.
我想避免列表理解，而只剩下纯粹的numpy
I would like a way to trace back which i,j gave each result.
Is this doable in a general way, or does it require the specific calculation of foo?
如果是这样，如何？
