# Cannot assign 'torch.nn.modules.container.Sequential' as parameter

**URL:** <https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388>\
**Category:** Uncategorized\
**Created:** [August 2, 2019, 4:47am UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388 "2019-08-02T04:47:38Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![aerinykim](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/aerinykim/32/5158_2.png) [@aerinykim](https://discuss.pytorch.org/u/aerinykim)\
**Post date:** [August 2, 2019, 4:47am UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388/1 "2019-08-02T04:47:39Z")

</div>

I was following this method  
([Dynamic parameter declaration in forward function](https://discuss.pytorch.org/t/dynamic-parameter-declaration-in-forward-function/427)) to dynamically assign parameters in `forward` function.

However, my parameter is not just one single weight tensor but it is `nn.Sequential` .

When I implement below:

```auto
class MyModule(nn.Module):
    def __init__ (self):
        # you need to register the parameter names earlier
        self.register_parameter('W_di', None)

    def forward(self, input):
        if self.W_di is None:
            self.W_di = nn.Sequential(
                nn.Linear(mL_n * 2, 1024),
                nn.ReLU(),
                nn.Linear(1024, self.hS)).to(device)

```

I get the following error.

```auto
 TypeError: cannot assign 'torch.nn.modules.container.Sequential' as parameter 'W_di' (torch.nn.Parameter or None expected)

```

Is there any way that I can register `nn.Sequential` as a whole param? Thanks!

---

<div class="post-metadata">

**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\
**Post date:** [August 4, 2019, 9:51pm UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388/2 "2019-08-04T21:51:15Z")

</div>

You could register the `nn.Sequential` container directly using `self.W_di = nn.Sequential(...)`, which will thus register all internal parameters.  
I’m not sure, if I understand the use case correctly, but if you need to get only this subset of parameters, you could call:

```python
model.W_di.parameters()

```

Would that work for you or do you need to handle these parameters somehow differently?

---

<div class="post-metadata">

**Author:** ![aerinykim](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/aerinykim/32/5158_2.png) [@aerinykim](https://discuss.pytorch.org/u/aerinykim)\
**Post date:** [August 5, 2019, 7:12pm UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388/3 "2019-08-05T19:12:01Z")

</div>

Hi ptrblck, thanks for the reply.  
As you see in the `forward` function `self.W_di = nn.Sequential(...)` is how I assigned `nn.Sequential` to `self.W_di`. However, when the `forward` is called, I’m getting an error saying

`TypeError: cannot assign 'torch.nn.modules.container.Sequential' as parameter 'W_di' (torch.nn.Parameter or None expected)`

I’m not trying to call parameters nor a subset of params. Just trying to run the training but got this error.

pytorch 1.1.0

---

<div class="post-metadata">

**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\
**Post date:** [August 5, 2019, 7:40pm UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388/4 "2019-08-05T19:40:06Z")

</div>

In your ` __init__ ` you are using:

```python
self.register_parameter('W_di', None)

```

which will create this error later.  
If you don’t need `W_di` as an `nn.Parameter`, you could just remove it.

---

<div class="post-metadata">

**Author:** ![aerinykim](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/aerinykim/32/5158_2.png) [@aerinykim](https://discuss.pytorch.org/u/aerinykim)\
**Post date:** [August 5, 2019, 9:35pm UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388/5 "2019-08-05T21:35:19Z")

</div>

But what if I need to declare it as a parameter? Maybe I wasn’t explaining it fully here but I’m trying to assign parameter in `foward` function following this approach [Dynamic parameter declaration in forward function](https://discuss.pytorch.org/t/dynamic-parameter-declaration-in-forward-function/427)

In that post, the Chief Crazy Person suggested using `self.register_parameter`.

---

<div class="post-metadata">

**Author:** ![ptrblck](https://discuss.pytorch.org/user_avatar/discuss.pytorch.org/ptrblck/32/1823_2.png) [@ptrblck](https://discuss.pytorch.org/u/ptrblck)\
**Post date:** [August 5, 2019, 10:17pm UTC](https://discuss.pytorch.org/t/cannot-assign-torch-nn-modules-container-sequential-as-parameter/52388/6 "2019-08-05T22:17:44Z")

</div>

Ah OK, I clearly misunderstood the use case.  
This code should work as the suggested one by Adam 😉

```python
class MyModel(nn.Module):
    def __init__ (self):
        super(MyModel, self). __init__ ()
        self.submodule = None
        
    def forward(self, x):
        if self.submodule is None:
            self.submodule = nn.Sequential(
                nn.Linear(1, 1),
                nn.ReLU()
            )
        x = self.submodule(x)
        return x

model = MyModel()
model(torch.randn(1, 1))
print(dict(model.named_parameters()))
> {'submodule.0.weight': Parameter containing:
tensor([[-0.1282]], requires_grad=True), 'submodule.0.bias': Parameter containing:
tensor([-0.7143], requires_grad=True)}

```