Deep Learning

Build each major neural architecture yourself

A guided tour of the field's main neural architectures in which you implement each one rather than just read about it: plain networks first, then convolutional models for vision, sequence models and Transformers for language, and finally generative models that synthesise new images.

What you will learn

→Train and tune feedforward neural networks in PyTorch
→Build convolutional networks for image classification and transfer learning
→Model sequences and text with RNNs and Transformer architectures
→Generate new images with GANs and diffusion models

Curriculum

1Constructing and training neural networks
2Building convolutional neural networks for computer vision
3Creating sequence models and Transformers
4Building generative models
GANs and diffusion models

Where an introductory course explains architectures, this one makes you implement them. You move from the feedforward basics into convolutional networks for computer vision, then sequence models and Transformers, and close on generative models: the GANs and diffusion systems behind synthetic imagery. It sits squarely inside deep learning, a deliberate step beyond general machine learning.

Prerequisites

Intermediate Python
NumPy, pandas and Matplotlib
Linear algebra (vectors, matrix operations) and basic calculus
Familiarity with feedforward neural networks and PyTorch

Concepts this course teaches

Deep Learning
The program's central subject, end to end.
Neural Networks
Builds and trains networks from the ground up.
Transformers
A full module on sequence models and Transformers.
Computer Vision
Convolutional networks for image tasks.
Generative AI
Builds GANs and diffusion models.
Natural Language Processing
Sequence models applied to language.

FAQ

How is this different from AI Programming with Python?: That program gets you to your first model; this one assumes you are already there and drills into the architectures (convolutional, sequence, Transformer and generative) one project at a time.
Will I build generative models myself?: Yes. The final stretch builds GANs and diffusion models, including a project that generates images, so you finish having shipped generative work rather than only reading about it.

Course facts

Provider: Udacity
Level: intermediate
Duration: About 50 hours · Self-paced
Price: Subscription: pricing set by Udacity
Pace: self-paced
Credential: nanodegree

PyTorchConvolutional neural networksRecurrent neural networksTransformersGenerative adversarial networksDiffusion modelsBackpropagationTransfer learning

View course at Udacity

Affiliate disclosure: course links are affiliate links; an enrolment through one may pay Getting Digital a commission from the provider, at no extra cost to the learner, and it does not influence how courses are described.