---
title: "A Simple Framework for Contrastive Learning of Visual Representations"
description: "SimCLR turns two altered views of one photograph into a training target without class labels. Read it to see why easy visual shortcuts can spoil representation learning, why the projection head is discarded after training, and why the headline ImageNet result needs its model size attached."
canonical_url: "https://fanout.sh/daily/2026-10-04-simclr-contrastive-representations"
md_url: "https://fanout.sh/daily/2026-10-04-simclr-contrastive-representations.md"
last_updated: "2026-10-04"
access: "public"
---

# A Simple Framework for Contrastive Learning of Visual Representations

SimCLR turns two altered views of one photograph into a training target without class labels. Read it to see why easy visual shortcuts can spoil representation learning, why the projection head is discarded after training, and why the headline ImageNet result needs its model size attached.

## Paper details

- Authors: Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton

- Venue: ICML 2020

- Track: AI Research

- Difficulty: Easy

- Reading time: 30 min

- Original paper: https://arxiv.org/abs/2002.05709v3

## What you will learn

- How same-source image views define positives while other batch views supply competing negatives.

- Why augmentation composition and a nonlinear projection head change the representation that transfers.

- How to interpret linear evaluation, batch-size tradeoffs, and the wider model behind the headline accuracy.

## Continue exploring

[Compare image-text pairing in Fanout](https://fanout.sh/daily/2026-08-23-clip-natural-language-supervision): Continue from matching two views of an image to CLIP’s matching of observed image-text pairs.

---
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