CIFAR-10
CIFAR-10 is a labeled image-classification dataset containing 60,000 32-by-32 color images from ten classes, divided into training and test sets.
Build a Text-to-Image Generator (from Scratch)
CIFAR-10 is a labeled image dataset containing color images divided into 10 classes for machine-learning evaluation and training.
- How it is used
- Supports training and evaluating computer-vision models, especially a vision transformer for image classification.
- How deep it goes
- Defines the dataset, download split, labels, visualization, dimensions, and ViT patch conversion without detailing dataset internals.
- Examples
- Uses planes, cars, birds, cats, frogs, and other named classes, plus plotted 3 × 6 grids of training and test images.
- What you can do after
- Operate/configure torchvision to download, split, label, batch, visualize, and evaluate CIFAR-10 images.
Read how Build a Text-to-Image Generator (from Scratch) explains CIFAR-10 →
Deep Learning with PyTorch, Second Edition
A benchmark dataset of labeled natural images.
- How it is used
- Supports hands-on image-classification exercises involving loading, preprocessing, batching, and convolutional modeling.
- How deep it goes
- Descends from downloading images through tensor conversion, channel-statistic computation, normalization, batching, and convolutional input shapes.
- Examples
- Uses 32 × 32 RGB images, per-channel means and standard deviations, normalized image plots, and a 3-to-16-channel Conv2d example.
- What you can do after
- Apply it via torchvision and PyTorch to load, normalize, batch, and classify labeled images with neural networks.
Read how Deep Learning with PyTorch, Second Edition explains CIFAR-10 →
Related concepts
- torchvisionused by · torchvision is a PyTorch library providing computer-vision datasets, models, image transformations, and related utilities.
- Vision Transformeruses · A Vision Transformer is a Transformer-based neural-network architecture that processes images as sequences of tokenized regions using attention.
- datasetkind of · A dataset is an organized collection of related data items, records, or observations treated as a unit for storage, processing, analysis, or computation.
- trainingused by · Training is a process of developing or improving a person’s or model’s capabilities through instruction, experience, or adjustment using data.
- RGBuses · RGB is an additive color model that represents colors through separate red, green, and blue component values or intensities.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.