Deep Learning (DL) · Intermediate level

Intermediate level · Module 1 of 6 · 15 guided hours

PyTorch in Depth

Write well-structured PyTorch code: tensors and autograd, custom datasets and loaders, and reusable training loops with checkpoints.

Learning objectives

  • Use tensors, devices and autograd confidently
  • Write custom Dataset classes and use DataLoader
  • Structure training code with configuration, logging and checkpoints

Topics

  • tensors, shapes and devices
  • autograd and computational graphs
  • Dataset and DataLoader
  • transforms
  • training and evaluation loops
  • checkpoints and resuming

Lessons

  1. Lesson 1 Tensors, Devices and Autograd

    Master tensor operations, move work between CPU and GPU, and understand how autograd computes gradients.

    tensor shape and dtypedevice (CPU/GPU)autogradtorch.no_grad

  2. Lesson 2 Datasets and DataLoaders

    Load data from folders or files with custom Dataset classes and efficient DataLoaders.

    Dataset __len__ and __getitem__DataLoadertransforms pipelineshuffling

  3. Lesson 3 Structuring Training Code

    Organise training into reusable functions with configuration, logging and checkpoints.

    train and evaluate functionsconfiguration filelogging metricscheckpoint (model + optimiser + epoch)

Practical activity

Refactor a notebook-only training script into a small project with a dataset class, a training module, a configuration file and checkpointing.

Expected outcomes

  • Reusable, debuggable training code
  • Reliable long training runs

Assessment

Code review against a checklist plus a demonstration of resuming training from a checkpoint.

Taught live by an instructor — online or in person. Free course lessons are for self-study and do not lead to a certificate — see registration options.