Hi, all.

Is there an efficient way to compute column-wise L1 norm between two matrices?

Input: x, N x d

y, M x d

Output: dist, N x M. Where dist[i,j] is the abs between x[i,:] and y[j,:].

i.e. dist[i,j] = ||x[i,:]-y[j,:]||

Hi, all.

Is there an efficient way to compute column-wise L1 norm between two matrices?

Input: x, N x d

y, M x d

Output: dist, N x M. Where dist[i,j] is the abs between x[i,:] and y[j,:].

i.e. dist[i,j] = ||x[i,:]-y[j,:]||