Function Differences with torch.nn.SequentialCell

torch.nn.Sequential

torch.nn.Sequential(
    *args
)

For more information, see torch.nn.Sequential.

mindspore.nn.SequentialCell

mindspore.nn.SequentialCell(
    *args
)

For more information, see mindspore.nn.SequentialCell.

Differences

PyTorch: Construct the Cell order container. Sequential adds the Cells in the order of the incoming List. In addition, OrderedDict is also supported as a constructor.

MindSpore: Construct the Cell order container. The input types are the same as PyTorch. In contrast to PyTorch, MindSpore supports append(), which adds the Cell at the end of the container.

Categories	Subcategories	PyTorch	MindSpore	Difference
Parameter	Parameter 1	args	args	Parameters of the incoming container, supporting List and OrderedDict types.

Code Example

import collections

# In MindSpore
import mindspore as ms

model = ms.nn.SequentialCell(
          ms.nn.Conv2d(1,20,5),
          ms.nn.ReLU(),
          ms.nn.Conv2d(20,64,5),
          ms.nn.ReLU()
        )
print(model)
# Out:
# SequentialCell<
#   (0): Conv2d<input_channels=1, output_channels=20, kernel_size=(5, 5), stride=(1, 1), pad_mode=same, padding=0, dilation=(1, 1), group=1, has_bias=False, weight_init=normal, bias_init=zeros, format=NCHW>
#   (1): ReLU<>
#   (2): Conv2d<input_channels=20, output_channels=64, kernel_size=(5, 5), stride=(1, 1), pad_mode=same, padding=0, dilation=(1, 1), group=1, has_bias=False, weight_init=normal, bias_init=zeros, format=NCHW>
#   (3): ReLU<>
#   >

# Example of using Sequential with OrderedDict
model = ms.nn.SequentialCell(collections.OrderedDict([
          ('conv1', ms.nn.Conv2d(1,20,5)),
          ('relu1', ms.nn.ReLU()),
          ('conv2', ms.nn.Conv2d(20,64,5)),
          ('relu2', ms.nn.ReLU())
        ]))
print(model)
# Out:
# SequentialCell<
#   (conv1): Conv2d<input_channels=1, output_channels=20, kernel_size=(5, 5), stride=(1, 1), pad_mode=same, padding=0, dilation=(1, 1), group=1, has_bias=False, weight_init=normal, bias_init=zeros, format=NCHW>
#   (relu1): ReLU<>
#   (conv2): Conv2d<input_channels=20, output_channels=64, kernel_size=(5, 5), stride=(1, 1), pad_mode=same, padding=0, dilation=(1, 1), group=1, has_bias=False, weight_init=normal, bias_init=zeros, format=NCHW>
#   (relu2): ReLU<>
#   >


# In PyTorch
import torch

model = torch.nn.Sequential(
          torch.nn.Conv2d(1,20,5),
          torch.nn.ReLU(),
          torch.nn.Conv2d(20,64,5),
          torch.nn.ReLU()
        )
print(model)
# Out
# Sequential(
#   (0): Conv2d(1, 20, kernel_size=(5, 5), stride=(1, 1))
#   (1): ReLU()
#   (2): Conv2d(20, 64, kernel_size=(5, 5), stride=(1, 1))
#   (3): ReLU()
# )

# Example of using Sequential with OrderedDict
model = torch.nn.Sequential(collections.OrderedDict([
          ('conv1', torch.nn.Conv2d(1,20,5)),
          ('relu1', torch.nn.ReLU()),
          ('conv2', torch.nn.Conv2d(20,64,5)),
          ('relu2', torch.nn.ReLU())
        ]))
print(model)
# Out：
# Sequential(
#   (conv1): Conv2d(1, 20, kernel_size=(5, 5), stride=(1, 1))
#   (relu1): ReLU()
#   (conv2): Conv2d(20, 64, kernel_size=(5, 5), stride=(1, 1))
#   (relu2): ReLU()
# )