Differences with torch.Tensor.repeat

View Source On Gitee

The following mapping relationships can be found in this file.

PyTorch APIs

MindSpore APIs

torch.Tensor.repeat

mindspore.Tensor.tile

torch.Tensor.repeat

torch.Tensor.repeat(*sizes) -> Tensor

For more information, see torch.Tensor.repeat.

mindspore.Tensor.tile

mindspore.Tensor.tile(multiples)

For more information, see mindspore.Tensor.tile.

Differences

The usage of mindspore.Tensor.tile is basically the same with that of torch.Tensor.repeat.

Categories

Subcategories

PyTorch

MindSpore

Differences

Parameters

Parameter 1

*sizes

multiples

In PyTorch, parameter type is torch.Size or int and in MindSpore, parameter type must be tuple.

Code Example

# PyTorch
import torch

input = torch.tensor([[1, 2], [3, 4]], dtype=torch.float32)
output = input.repeat(16, 1)
print(output.shape)
# torch.Size([32, 2])

# MindSpore
import mindspore

x = mindspore.Tensor([[1, 2], [3, 4]], dtype=mindspore.float32)
output = x.tile((16, 1))
print(output.shape)
# (32, 2)