Differences with torch.renorm

The following mapping relationships can be found in this file.

PyTorch APIs	MindSpore APIs
torch.renorm	mindspore.ops.renorm
torch.Tensor.renorm	mindspore.Tensor.renorm

torch.renorm

torch.renorm(input, p, dim, maxnorm, *, out=None) -> Tensor

For more information, see torch.renorm.

mindspore.ops.renorm

mindspore.ops.renorm(input, p, axis, maxnorm)

For more information, see mindspore.ops.renorm.

Differences

API function of MindSpore is consistent with that of PyTorch.

PyTorch: The data type of parameter p is float .

MindSpore: The data type of parameter p is int .

Categories	Subcategories	PyTorch	MindSpore	Differences
Parameters	Parameter 1	input	input	-
	Parameter 2	p	p	The data type supported by PyTorch is float , the data type supported by MindSpore is int .
	Parameter 3	dim	axis	Different parameter names
	Parameter 4	maxnorm	maxnorm	-
	Parameter 5	out	-	For details, see General Difference Parameter Table

Code Example 1

# PyTorch
import torch
x = torch.tensor([[0, 1, 2, 3], [4, 5, 6, 7], [8, 9, 10, 11]], dtype=torch.float32)
out = torch.renorm(x, 2.0, 0, 5.0)
print(out.numpy())
# [[0.        1.        2.        3.       ]
#  [1.7817416 2.2271771 2.6726124 3.1180477]
#  [2.0908334 2.3521876 2.6135418 2.874896 ]]


# MindSpore
import mindspore
import mindspore.ops as ops
from mindspore import Tensor

x = Tensor([[0, 1, 2, 3], [4, 5, 6, 7], [8, 9, 10, 11]], dtype=mindspore.float32)
out = ops.renorm(x, 2, 0, 5.0)
print(out.numpy())
# [[0.        1.        2.        3.       ]
#  [1.7817416 2.2271771 2.6726124 3.118048 ]
#  [2.0908334 2.3521876 2.6135418 2.874896 ]]