Function Differences with torch.nn.CrossEntropyLoss

class torch.nn.CrossEntropyLoss(
    weight=None,
    size_average=None,
    ignore_index=-100,
    reduce=None,
    reduction='mean'
)(input, target) -> Tensor

class mindspore.nn.CrossEntropyLoss(
    weight=None,
    ignore_index=-100,
    reduction='mean',
    label_smoothing=0.0
)(logits, labels) -> Tensor

# PyTorch
import torch
import numpy as np

input_torch = np.array([[1.624, -0.611, -0.528, -1.072, 0.865], [-2.301, 1.744, -0.761, 0.319, -0.249], [1.462, -2.060, -0.322, -0.384, 1.133]])
target_torch = np.array([1, 0, 4])
loss = torch.nn.CrossEntropyLoss()
input_torch = torch.tensor(input_torch, requires_grad=True)
target_torch = torch.tensor(target_torch, dtype=torch.long)
output = loss(input_torch, target_torch)
print(round(float(output.detach().numpy()), 3))
# 2.764

# MindSpore
import mindspore
import numpy as np

input_ms = np.array([[1.624, -0.611, -0.528, -1.072, 0.865], [-2.301, 1.744, -0.761, 0.319, -0.249], [1.462, -2.060, -0.322, -0.384, 1.133]])
target_ms = np.array([1, 0, 4])
input_ms = mindspore.Tensor(input_ms, mindspore.float32)
target_ms = mindspore.Tensor(target_ms, mindspore.int32)
loss = mindspore.nn.CrossEntropyLoss()
output = loss(input_ms, target_ms)
print(round(float(output), 3))
# 2.764