Home › Research Papers › AlexNet (2012)
AlexNet (2012)
Research Paper
Architecture
The ImageNet result that started the deep-learning era — on two gaming GPUs.
You're looking at a research-paper record: a landmark paper and what it introduced to the field.
Krizhevsky, Sutskever and Hinton halved the ImageNet error rate with a convolutional network trained on two consumer NVIDIA GTX 580s — the moment that proved GPUs + data + depth beats hand-engineering, redirected the field, and incidentally set NVIDIA on its current path. Every GPU cluster in this database descends from that dorm-room-scale rig.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Training data includes
Authored by
Related
Sources
Record reviewed as of 2026-01-15 — see Methodology.