mindspore.Tensor.bincount

Tensor.bincount(weights=None, minlength=0) → Tensor

Count the occurrences of each value in the self.

If minlength is not specified, the length of the output Tensor is the maximum value in the self plus one. If minlength is specified, the length of the output Tensor is the maximum value between minlength and the maximum value in the self plus one.

Each value in the output Tensor represents the number of occurrences of that index value in the self. If weights is specified, the output results are weighted, i.e., \(out[n] += weight[i]\) instead of \(out[n] += 1\).

Parameters

  • weights (Tensor, optional) – Weights with the same shape as the self . Default: None .

  • minlength (int, optional) – The minimum length of output Tensor. Should be non-negative. Default: 0 .

Returns

Tensor, If self is non-empty, the output shape is \((max(max(self)+1, minlength), )\), otherwise the shape is \((0, )\).

Raises

  • TypeError – If weights is not a Tensor.

  • ValueError – If self contains negative values.

  • ValueError – If self is not one-dimensional or self and weights do not have the same shape.

Supported Platforms:

Ascend GPU CPU

Examples

>>> from mindspore import Tensor
>>> from mindspore import dtype as mstype
>>> x = Tensor([2, 4, 1, 0, 0], dtype=mstype.int64)
>>> print(ops.bincount(x, minlength=7))
[2. 1. 1. 0. 1. 0. 0.]
>>> weights = Tensor([0, 0.25, 0.5, 0.75, 1], dtype=mstype.float32)
>>> print(x.bincount(weights=weights))
[1.75 0.5  0.   0.   0.25]