LINEAiGE
AI HISTORY

Caffe Technical Paper

The Caffe authors recorded the framework in the paper 'Caffe: Convolutional Architecture for Fast Feature Embedding' on June 20, 2014, describing Caffe as a fully open-source deep-learning framework implemented in C++ with CUDA and bindings for Python and MATLAB.

20 June 2014AI HISTORYVERIFIED

Historical classification

This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.

Why it matters

A primary-source, day-precision documentation milestone for Caffe that records the framework without converting the paper date into an unsupported invention, system-birth, or public-release date.

People, organizations & technologies

PEOPLE
Yangqing Jia, Evan Shelhamer, Jeff Donahue, Sergey Karayev, Jonathan Long, Ross Girshick, Sergio Guadarrama, Trevor Darrell

ORGANIZATIONS
Berkeley Vision and Learning Center

TECHNOLOGIES
Caffe

Evidence & verification

STATE
verified

CONFIDENCE
high

LAST VERIFIED
2026-09-13T11:03:07+08:00

The authors' primary paper gives a June 20, 2014 date and directly describes Caffe as a fully open-source framework, its C++/CUDA implementation, Python and MATLAB bindings, and BVLC maintenance/development. The official Caffe project page independently corroborates Yangqing Jia as creator, Evan Shelhamer as lead developer, BAIR/community development, and BSD 2-Clause licensing. This candidate records only the paper milestone and source-explicit organizational attribution; it does not infer an invention/system-birth date, exact public-release date, development geography, universal priority, ancestry edge, influence, causality, or other unsupported relationship.

Sources

SOURCE 1 · PRIMARY · CLAIM-EVIDENCE

Caffe: Convolutional Architecture for Fast Feature Embedding

arXiv

PUBLISHED
2014-06-20

SOURCE 2 · PRIMARY · CORROBORATION

Caffe | Deep Learning Framework

Berkeley AI Research

Trace the chronology

Record identity

LINEAiGE ID
caffe-paper-2014