Differences with torch.nn.RNNCell

torch.nn.RNNCell

class torch.nn.RNNCell(
    input_size,
    hidden_size,
    bias=True,
    nonlinearity='tanh')(input, hidden) -> Tensor

For more information, see torch.nn.RNNCell.

mindspore.nn.RNNCell

class mindspore.nn.RNNCell(
    input_size: int,
    hidden_size: int,
    has_bias: bool=True,
    nonlinearity: str = 'tanh')(x, hx) -> Tensor

For more information, see mindspore.nn.RNNCell.

Differences

PyTorch: Recurrent Neural Network (RNN) unit.

MindSpore: MindSpore API implements the same functions as PyTorch.

Categories	Subcategories	PyTorch	MindSpore	Difference
Parameters	Parameter 1	input_size	input_size	-
	Parameter 2	hidden_size	hidden_size	-
	Parameter 3	bias	has_bias	Same function, different parameter names
	Parameter 4	nonlinearity	nonlinearity	-
Inputs	Input 1	input	x	Same function, different parameter names
	Input 2	hidden	hx	Same function, different parameter names

Code Example 1

# PyTorch
import torch
from torch import tensor
import numpy as np

rnncell = torch.nn.RNNCell(2, 3, nonlinearity="relu", bias=False)
input = torch.tensor(np.array([[3.0, 4.0]]).astype(np.float32))
hidden = torch.tensor(np.array([[1.0, 2.0, 3]]).astype(np.float32))
output = rnncell(input, hidden)
print(output)
# tensor([[0.5022, 0.0000, 1.4989]], grad_fn=<ReluBackward0>)

# MindSpore
import mindspore.nn as nn
from mindspore import Tensor
import numpy as np

rnncell = nn.RNNCell(2, 3, nonlinearity="relu", has_bias=False)
x = Tensor(np.array([[3.0, 4.0]]).astype(np.float32))
hx = Tensor(np.array([[1.0, 2.0, 3]]).astype(np.float32))
output = rnncell(x, hx)
print(output)
# [[2.4998584 0.        1.9334991]]