mars.tensor.nanmax¶

mars.tensor.
nanmax
(a, axis=None, out=None, keepdims=None, combine_size=None)[source]¶ Return the maximum of an array or maximum along an axis, ignoring any NaNs. When allNaN slices are encountered a
RuntimeWarning
is raised and NaN is returned for that slice. Parameters
a (array_like) – Tensor containing numbers whose maximum is desired. If a is not a tensor, a conversion is attempted.
axis (int, optional) – Axis along which the maximum is computed. The default is to compute the maximum of the flattened tensor.
out (ndarray, optional) – Alternate output array in which to place the result. The default is
None
; if provided, it must have the same shape as the expected output, but the type will be cast if necessary. See doc.ufuncs for details.keepdims (bool, optional) –
If this is set to True, the axes which are reduced are left in the result as dimensions with size one. With this option, the result will broadcast correctly against the original a.
If the value is anything but the default, then keepdims will be passed through to the max method of subclasses of Tensor. If the subclasses methods does not implement keepdims any exceptions will be raised.
combine_size (int, optional) – The number of chunks to combine.
 Returns
nanmax – A tensor with the same shape as a, with the specified axis removed. If a is a 0d tensor, or if axis is None, a Tensor scalar is returned. The same dtype as a is returned.
 Return type
Tensor
See also
nanmin
The minimum value of a tensor along a given axis, ignoring any NaNs.
amax
The maximum value of a tensor along a given axis, propagating any NaNs.
fmax
Elementwise maximum of two tensors, ignoring any NaNs.
maximum
Elementwise maximum of two tensors, propagating any NaNs.
isnan
Shows which elements are Not a Number (NaN).
isfinite
Shows which elements are neither NaN nor infinity.
Notes
Mars uses the IEEE Standard for Binary FloatingPoint for Arithmetic (IEEE 754). This means that Not a Number is not equivalent to infinity. Positive infinity is treated as a very large number and negative infinity is treated as a very small (i.e. negative) number.
If the input has a integer type the function is equivalent to np.max.
Examples
>>> import mars.tensor as mt
>>> a = mt.array([[1, 2], [3, mt.nan]]) >>> mt.nanmax(a).execute() 3.0 >>> mt.nanmax(a, axis=0).execute() array([ 3., 2.]) >>> mt.nanmax(a, axis=1).execute() array([ 2., 3.])
When positive infinity and negative infinity are present:
>>> mt.nanmax([1, 2, mt.nan, mt.NINF]).execute() 2.0 >>> mt.nanmax([1, 2, mt.nan, mt.inf]).execute() inf