Learn and practice PyTorch, a framework for developing deep learning models.

Become a Deep Learning Professional

In this module, we’ll be introduced to the fundamentals of PyTorch: autograd, model classes, datasets, data loaders, and more. We’ll develop models and understand them step-by-step. We'll understand the reasoning behind the code and how to avoid some common pitfalls and errors along the way.
By the time we finish this module, we’ll have a thorough understanding of the concepts and tools necessary to start developing and training our own models using PyTorch.

Deep Learning with PyTorch

Check all that apply. Which methods should be implemented in a custom model? 

Check all that apply. Which methods should be implemented in a custom model?

Given the partially implemented models below, how many tensors are returned by `MyModel().parameters()`?

```python
class SubModel(nn.Module):
  def __init__(self):
    super().__init__()
    self.w1 = nn.Parameter(torch.randn(1, dtype=torch.float))
    self.w2 = nn.Parameter(torch.randn(1, dtype=torch.float))

class MyModel(nn.Module):
  def __init__(self):
    super().__init__()
    self.linear = nn.Linear(1, 1, bias=False)
    self.sub = SubModel()
```

Given the partially implemented models below, how many tensors are returned by <code>MyModel().parameters()</code>?
<pre><code class="language-python">class SubModel(nn.Module):
 def __init__(self):
 super().__init__()
 self.w1 = nn.Parameter(torch.randn(1, dtype=torch.float))
 self.w2 = nn.Parameter(torch.randn(1, dtype=torch.float))

class MyModel(nn.Module):
 def __init__(self):
 super().__init__()
 self.linear = nn.Linear(1, 1, bias=False)
 self.sub = SubModel()
</code></pre>

What's the recommended way of performing the forward pass in a model?

What’s the recommended way of performing the forward pass in a model?

Given the snippet below, what is the output produced by calling `model(x)`?

```python
x = torch.randn(10, 1)
model = nn.Sequential(nn.Linear(1, 1))
model.to('cuda')
```

Given the snippet below, what is the output produced by calling <code>model(x)</code>?
<pre><code class="language-python">x = torch.randn(10, 1)
model = nn.Sequential(nn.Linear(1, 1))
model.to('cuda')
</code></pre>

What's the purpose of calling `model.train()`?

What’s the purpose of calling <code>model.train()</code>?

Check all that apply. Which elements have a state dictionary? 

Check all that apply. Which elements have a state dictionary?

Given the sequential linear model below, how many tensors are returned by `model.parameters()`?

```python
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = nn.Sequential(nn.Linear(1, 1)).to(device)
```

Given the sequential linear model below, how many tensors are returned by <code>model.parameters()</code>?
<pre><code class="language-python">device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = nn.Sequential(nn.Linear(1, 1)).to(device)
</code></pre>

1) `model.train()`
2) `optimizer.step()`
3) `loss.backward()`

1) `loss.backward()`
2) `optimizer.step()`
3) `model.train()`

1) `model.train()`
2) `loss.backward()`
3) `optimizer.step()`

1) `optimizer.step()`
2) `loss.backward()`
3) `model.train()`

You have been provided with the following code snippet:

```python
torch.manual_seed(42)
model = nn.Sequential(nn.Linear(1, 1)).to(device)        
optimizer = optim.SGD(model.parameters(), lr=lr)
loss_fn = nn.MSELoss(reduction='mean')

for epoch in range(n_epochs):
    1) ....                

    yhat = model(x_train_tensor)                   
    
    loss = loss_fn(yhat, y_train_tensor)

    2) ....
    
    3) ....
    optimizer.zero_grad()
```

Using the choices given below, fill in the blanks by selecting the correct options at their respective places:

You have been provided with the following code snippet:
<pre><code class="language-python">torch.manual_seed(42)
model = nn.Sequential(nn.Linear(1, 1)).to(device) 
optimizer = optim.SGD(model.parameters(), lr=lr)
loss_fn = nn.MSELoss(reduction='mean')

for epoch in range(n_epochs):
 1) .... 

 yhat = model(x_train_tensor) 
 
 loss = loss_fn(yhat, y_train_tensor)

 2) ....
 
 3) ....
 optimizer.zero_grad()
</code></pre>
Using the choices given below, fill in the blanks by selecting the correct options at their respective places:

Test your understanding of the concepts explained in this chapter with this quiz!

Introduction

Visualizing Gradient Descent

A Simple Regression Problem

Rethinking the Training Loop

Going Classy

A Simple Classification Problem

Appendix

Conclusion

Quiz