Deep Learning with PyTorch
Build neural networks for vision, NLP, and generative AI using PyTorch.
About this course
A comprehensive deep-learning curriculum using PyTorch. You will implement feedforward networks, CNNs, RNNs, transformers, and diffusion models from scratch, train them on real datasets, and learn the engineering practices needed to ship models to production. GPU-accelerated notebooks are provided.
Target audience: ML engineers, researchers, advanced students
What you will learn
- PyTorch
- CNNs
- Transformers
- Transfer learning
- Model serving
Course syllabus
10 modules · video + projects
- 1PyTorch tensors and the autograd system
- 2Feedforward networks and backpropagation
- 3Convolutional neural networks for image classification
- 4Transfer learning and fine-tuning pre-trained models
- 5Recurrent networks and LSTMs for sequence data
- 6The transformer architecture from scratch
- 7Fine-tuning large language models
- 8Generative models: VAEs and diffusion
- 9Training at scale: mixed precision, gradient checkpointing
- 10Serving a model with FastAPI and Docker
Prerequisites
- –Python
- –Machine Learning Fundamentals or equivalent
- –Linear algebra basics
Frequently asked questions
Do I need a GPU?
No. All notebooks run on GeraLearn's cloud GPU environment included with your enrolment. You can also run locally with a CPU for the smaller exercises.
Is this course up to date with 2026 techniques?
Yes. The curriculum was updated in Q1 2026 to include modern transformer fine-tuning workflows and diffusion model basics.
Ready to start Deep Learning with PyTorch?
Join 3,200+ learners already enrolled. Self-paced, certificate on completion.