Differences with torch.Tensor.flip

torch.Tensor.flip

torch.Tensor.flip(dims)

For more details, see torch.Tensor.flip.

mindspore.Tensor.flip

mindspore.Tensor.flip(dims)

For more details, see mindspore.Tensor.flip.

Differences

PyTorch: The torch.Tensor.flip interface has differences from torch.flip. Compared to torch.flip, Tensor.flip additionally supports scenarios where the dims input is of type int.

MindSpore: The mindspore.flip and mindspore.Tensor.flip interfaces have the same functionality as torch.flip and do not support input of type int.

Categories	Subcategories	PyTorch	MindSpore	Differences
Parameters	Parameter 1	dims	dims	Same functionality, MindSpore does not support int input

Code Example

# PyTorch
import numpy as np
import torch
input = torch.tensor(np.arange(1, 9).reshape((2, 2, 2)))
output = input.flip(1)
print(output)
# tensor([[[3, 4],
#          [1, 2]],
#
#         [[7, 8],
#          [5, 6]]])

# MindSpore
import mindspore as ms
import mindspore.ops as ops

input = ms.Tensor(np.arange(1, 9).reshape((2, 2, 2)))
output = input.flip((1, ))
print(output)
# [[[3 4]
#   [1 2]]
#
#  [[7 8]
#   [5 6]]]