Self-driving cars, smartphones, search engines... Deep learning is now everywhere. Before you begin building complex models, you will become familiar with PyTorch, a deep learning framework. You will learn how to manipulate tensors, create PyTorch data structures, and build your first neural network in PyTorch with linear layers.

Introduction to deep learning with PyTorch

Getting started with PyTorch tensors

Checking and adding tensors

Neural networks and layers

Linear layer network

Understanding weights

Hidden layers and parameters

Your first neural network

Stacking linear layers

Counting the number of parameters

Introduction to PyTorch, a Deep Learning Library

To train a neural network in PyTorch, you will first need to understand additional components, such as activation and loss functions. You will then realize that training a network requires minimizing that loss function, which is done by calculating gradients. You will learn how to use these gradients to update your model's parameters.


Discovering activation functions

Activate your understanding!

The sigmoid and softmax functions

Running a forward pass

Building a binary classifier in PyTorch

From regression to multi-class classification

Using loss functions to assess model predictions

Creating one-hot encoded labels

Calculating cross entropy loss

Using derivatives to update model parameters

Accessing the model parameters

Updating the weights manually

Using the PyTorch optimizer

Neural Network Architecture and Hyperparameters

Now that you've learned the key components of a neural network, you'll train one using a training loop. You'll explore potential issues like vanishing gradients and learn strategies to address them, such as alternative activation functions and tuning learning rate and momentum.

A deeper dive into loading data

Using TensorDataset

Using DataLoader

Writing our first training loop

Using the MSELoss

Writing a training loop

ReLU activation functions

Implementing ReLU

Implementing leaky ReLU

Understanding activation functions

Learning rate and momentum

Experimenting with learning rate

Experimenting with momentum

Training a Neural Network with PyTorch

Training a deep learning model is an art, and to make sure our model is trained correctly, we need to keep track of certain metrics during training, such as the loss or the accuracy. We will learn how to calculate such metrics and how to reduce overfitting.


Layer initialization and transfer learning

Fine-tuning process

Freeze layers of a model

Layer initialization

Evaluating model performance

Writing the evaluation loop

Calculating accuracy using torchmetrics

Fighting overfitting

Experimenting with dropout

Understanding overfitting

Improving model performance

Implementing random search

Wrap-up video

Evaluating and Improving Models

Water Potability

Face Mask Dataset

Course Glossary

Deep learning is everywhere, from smartphone cameras to voice assistants or self-driving cars. In this course, you will discover this powerful technology and learn how to leverage it using PyTorch, one of the most popular deep learning libraries. By the end of this course, you will be able to leverage PyTorch to solve classification and regression problems using deep learning.

The videos contain live transcripts you can reveal by clicking "Show transcript" at the bottom left of the videos.

The course glossary can be found on the right in the resources section.

To obtain CPE credits you need to complete the course and reach a score of 70% on the qualified assessment. You can navigate to the assessment by clicking on the CPE credits callout on the right.

<h2>Understanding the power of Deep Learning</h2>
Deep learning is everywhere: in smartphone cameras, voice assistants, and self-driving cars. It has even helped discover protein structures and beat humans at the game of Go. Discover this powerful technology and learn how to leverage it using PyTorch, one of the most popular deep learning libraries.<br><br>
<h2>Train your first neural network</h2>First, tackle the difference between deep learning and "classic" machine learning. You will learn about the training process of a neural network and how to write a training loop. To do so, you will create loss functions for regression and classification problems and leverage PyTorch to calculate their derivatives.<br><br><h2>Evaluate and improve your model</h2>In the second half, learn the different hyperparameters you can adjust to improve your model. After learning about the different components of a neural network, you will be able to create larger and more complex architectures. To measure your model performances, you will leverage TorchMetrics, a PyTorch library for model evaluation. <br><br>Upon completion, you will be able to leverage PyTorch to solve classification and regression problems on both tabular and image data using deep learning. A vital capability for experienced data professionals looking to advance their careers.

Supervised Learning with scikit-learn

Introduction to NumPy

Python Toolbox

Learn how to build your first neural network, adjust hyperparameters, and tackle classification and regression problems in PyTorch.

Understanding activation functions

Introduction to Deep Learning with PyTorch

Esercizio pratico interattivo