Wikimedia: 40 Million Images for Computer Vision

Over 40 million images sourced from Wikimedia Commons comprise this collection curated by ryanrudes. Updated in October 2023, the repository provides a massive scale of visual data for deep learning and computer vision research.

Use Cases

• Training Generative Adversarial Networks (GANs) using the 40 million image files
• Self-supervised pre-training for computer vision using the diverse visual corpus
• Large-scale image classification research using Wikimedia-sourced media

Strengths

• Contains over 40 million individual images
• Sourced from the diverse Wikimedia Commons repository
• Released under the permissive MIT license

Limitations

• Lack of structured metadata or labels in the primary description
• Potential for significant class imbalance given the crowdsourced nature of Wikimedia
• High storage and bandwidth requirements for 40 million files

Provenance

Source

Wikimedia Commons

Collection Method

scraped

Freshness

Last updated October 2023; no scheduled update frequency provided.

Geography

global