---
title: "ImageNet Classification with Deep Convolutional Neural Networks"
description: "AlexNet asks how a deep image classifier can learn from 1.2 million examples without taking impossibly long to optimize or simply memorizing the training set. Read it to see how ReLU activations, a selectively connected two-GPU network, image augmentation, and dropout made an eight-layer convolutional model practical at ImageNet scale."
canonical_url: "https://fanout.sh/daily/2026-09-01-alexnet-image-classification"
md_url: "https://fanout.sh/daily/2026-09-01-alexnet-image-classification.md"
last_updated: "2026-09-01"
access: "public"
---

# ImageNet Classification with Deep Convolutional Neural Networks

AlexNet asks how a deep image classifier can learn from 1.2 million examples without taking impossibly long to optimize or simply memorizing the training set. Read it to see how ReLU activations, a selectively connected two-GPU network, image augmentation, and dropout made an eight-layer convolutional model practical at ImageNet scale.

## Paper details

- Authors: Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton

- Venue: NeurIPS 2012

- Track: AI Research

- Difficulty: Easy

- Reading time: 35 min

- Original paper: https://papers.nips.cc/paper_files/paper/2012/hash/c399862d3b9d6b76c8436e924a68c45b-Abstract.html

## What you will learn

- Why ReLU reached a fixed training-error threshold much sooner than tanh in the paper's controlled comparison, without implying one universal speedup.

- How two 3 GB GPUs divided the feature maps, exchanged information at selected layers, and converged into one 1,000-class classifier.

- How random crops, reflections, color perturbations, dropout, test-time crop averaging, and ensembles shape what the headline benchmark does and does not prove.

## Continue exploring

[Explore convolutional networks in Fanout](https://fanout.sh/knowledge-graph?node=ai-research%3Acategory%3Aglossary-architecture&detail=ai-research%3Aconcept%3Acnn): Connect AlexNet's local filters and shared weights to Fanout's convolutional-network concept, then trace the later vision architectures that changed depth, normalization, and scaling.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md