Function Differences with torch.empty

torch.empty

torch.empty(
    *size,
    *,
    out=None,
    dtype=None,
    layout=torch.strided,
    device=None,
    requires_grad=False,
    pin_memory=False,
    memory_format=torch.contiguous_format
) -> Tensor

For more information, see torch.empty.

mindspore.numpy.empty

mindspore.numpy.empty(shape, dtype=mstype.float32) -> Tensor

For more information, see mindspore.numpy.empty.

Differences

PyTorch: Return an uninitialized tensor, the shape of which is defined by size.

MindSpore: MindSpore API basically implements the same function as PyTorch, but the default value of the dtype parameter is different.

Code Example 1

For the parameter dtype, PyTorch defaults to None, and the output type is torch.float32, while MindSpore defaults to mstype.float32.

# PyTorch
import torch

torch_output = torch.empty(2, 3)
print(list(torch_output.shape))
# [2, 3]

# MindSpore
import mindspore

ms_output = mindspore.numpy.empty((2, 3))
print(ms_output.shape)
# (2, 3)