An issue in using custom C extensions

Hi all,

I try to write a C extension according to this: http://pytorch.org/tutorials/advanced/c_extension.html.
Everything is OK when I do the step 1. I can generate the _ext source folder and the .so file.
However, when I do the second step, something wrong happens.
My code is:

import torch
from _ext import my_lib
from torch.autograd import Function
from torch.nn import Module
import torch.nn as nn
from torch.autograd import Variable


class MyAddFunction(Function):
    def forward(self, input1, input2):
        output = torch.FloatTensor()
        my_lib.my_lib_add_forward(input1, input2, output)
        return output

    def backward(self, grad_output):
        grad_input = torch.FloatTensor()
        my_lib.my_lib_add_backward(grad_output, grad_input)
        return grad_input



class MyAddModule(Module):
    def forward(self, input1, input2):
        return MyAddFunction()(input1, input2)



class MyNetwork(nn.Module):
    def __init__(self):
        super(MyNetwork, self).__init__(
            add = MyAddModule(),
        )

    def forward(self, input1, input2):
        return self.add(input1, input2)



model = MyNetwork()
input1, input2 = Variable(torch.randn(5, 5)), Variable(torch.randn(5, 5))
print(model(input1, input2))
print(input1 + input2)

When I ran it, I got an error:

  File "./test.py", line 40, in <module>
    model = MyNetwork()
  File "./test.py", line 32, in __init__
    add = MyAddModule(),
TypeError: __init__() got an unexpected keyword argument 'add'

Could you tell me what’s wrong with it? Thanks!

this seems wrong

super(MyNetwork, self).__init__(
            add = MyAddModule(),
        )

Instead do:

super(MyNetwork, self).__init__():
self.add = MyAddModule()