Based on official python doc, the output of round function is integer when ndigits is omitted. However, numpy does not follow this rule, and the output of rounded numpy still is a float. I think this behavior cause bugs!

Reproducing code example:

>>> for x in np.linspace(0,1,5):
...     print(round(x * 5.))
... 
0.0
1.0
2.0
4.0
5.0
>>> for x in map(float,np.linspace(0,1,5)):
...     print(round(x * 5.))
... 
0
1
2
4
5

Numpy/Python version information:

1.18.1 3.5.2 (default, Oct 8 2019, 13:06:37)
[GCC 5.4.0 20160609]