Introduction to Deep Learning · HIT

Week 4   Part II · Training Infrastructure

Data Pipelines

The Dataset and DataLoader abstractions; batching, shuffling, transforms, and splits.

Curated, free, canonical references for this week: a course or lecture, a book chapter, a video, and an authoritative blog post or official tutorial. Each opens in a new tab.

Course
fast.ai Practical Deep Learning, Lesson 1: Getting Started course.fast.ai

Builds and trains an image model end-to-end with an automatic train/validation split.

Book
Dive into Deep Learning, 4.2 The Image Classification Dataset d2l.ai

Loads Fashion-MNIST and reads minibatches via a DataLoader, demonstrating batch_size and shuffling.

Video
PyTorch Tutorial: Dataset and DataLoader, Batch Training (Patrick Loeber) youtube.com

Focused walkthrough of writing a custom Dataset and wrapping it in a DataLoader.

Blog / Docs
PyTorch Official Tutorial: Datasets & DataLoaders docs.pytorch.org

The canonical tutorial on the Dataset/DataLoader primitives, batching, and shuffling.

← Back to the Week 4 lab

PreviousWeek 3: MLPs & BackpropagationNextWeek 5: Loss Functions & Metrics