mindspore.mint.scatter

mindspore.mint.scatter(input, dim, index, src)

Update the value in src to input according to the specified index. For a 3-D tensor, the output will be:

output[index[i][j][k]][j][k] = src[i][j][k]  # if dim == 0

output[i][index[i][j][k]][k] = src[i][j][k]  # if dim == 1

output[i][j][index[i][j][k]] = src[i][j][k]  # if dim == 2

Note

The backward propagation is supported only for the case src.shape == index.shape when src is a tensor. Each value of index must be in the valid range [0, input.shape[dim]) for the normalized dim. Negative or out-of-range index values are illegal input and will lead to undefined behavior.

Parameters:

• input (Tensor) – The target tensor. The rank of input must be at least 1.

• dim (int) – The axis along which to scatter. Accepted range is [-r, r) where r = rank(input).

• index (Tensor) – The index to do update operation whose data type must be mindspore.int32 or mindspore.int64. Same rank as input.

• src (Tensor, float) – The data doing the update operation with input. Can be a tensor with the same data type as input or a float number to scatter.

Returns:

Tensor, has the same shape and type as input .

Raises:

• TypeError – If index is neither int32 nor int64.

• ValueError – If rank of any of input , index and src is less than 1.

• ValueError – If the rank of src is not equal to the rank of input .

• TypeError – If the data types of input and src have different dtypes.

Supported Platforms:

Ascend

Examples

>>> import numpy as np
>>> import mindspore as ms
>>> from mindspore import Tensor, mint
>>> input = Tensor(np.array([[1, 2, 3, 4, 5]]), dtype=ms.float32)
>>> src = Tensor(np.array([[8, 8]]), dtype=ms.float32)
>>> index = Tensor(np.array([[2, 4]]), dtype=ms.int64)
>>> out = mint.scatter(input=input, dim=1, index=index, src=src)
>>> print(out)
[[1. 2. 8. 4. 8.]]
>>> input = Tensor(np.zeros((5, 5)), dtype=ms.float32)
>>> src = Tensor(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), dtype=ms.float32)
>>> index = Tensor(np.array([[0, 0, 0], [2, 2, 2], [4, 4, 4]]), dtype=ms.int64)
>>> out = mint.scatter(input=input, dim=0, index=index, src=src)
>>> print(out)
[[1. 2. 3. 0. 0.]
[0. 0. 0. 0. 0.]
[4. 5. 6. 0. 0.]
[0. 0. 0. 0. 0.]
[7. 8. 9. 0. 0.]]
>>> input = Tensor(np.zeros((5, 5)), dtype=ms.float32)
>>> src = Tensor(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), dtype=ms.float32)
>>> index = Tensor(np.array([[0, 2, 4], [0, 2, 4], [0, 2, 4]]), dtype=ms.int64)
>>> out = mint.scatter(input=input, dim=1, index=index, src=src)
>>> print(out)
[[1. 0. 2. 0. 3.]
[4. 0. 5. 0. 6.]
[7. 0. 8. 0. 9.]
[0. 0. 0. 0. 0.]
[0. 0. 0. 0. 0.]]