Pytorch Median - Is It Bug Or Am I Using It Wrong

I am trying to get median of each row of 2D torch.tensor. But the result is not what I expect when compared to working with standard array or numpy import torch import numpy as np

Solution 1:

Looks like this is the intended behaviour of Torch as mentioned in this issue

https://github.com/pytorch/pytorch/issues/1837
https://github.com/torch/torch7/pull/182

The reasoning as mentioned in the link above

Median returns 'middle' element in case of odd-many elements, otherwise one-before-middle element (could also do the other convention to take mean of the two around-the-middle elements, but that would be twice more expensive, so I decided for this one).

Solution 2:

You can emulate numpy median with pytorch:

import torch
import numpy as np
y =[1, 2, 3, 5, 9, 1]
print("numpy=",np.median(y))
print(sorted([1, 2, 3, 5, 9, 1]))
yt = torch.tensor(y,dtype=torch.float32)
ymax = torch.tensor([yt.max()])
print("torch=",yt.median())
print("torch_fixed=",(torch.cat((yt,ymax)).median()+yt.median())/2.)