Self-supervised Pretraining of Visual Features in the Wild

TL;DR

This paper demonstrates that large self-supervised models trained on uncurated, random images in real-world settings can achieve high accuracy, confirming the practical viability of self-supervised learning beyond curated datasets.

Contribution

The authors train a large-scale self-supervised model on uncurated images, showing it surpasses previous models and performs well in real-world scenarios.

Findings

SEER model achieves 84.2% top-1 accuracy on ImageNet.

Self-supervised models are effective few-shot learners.

Training on uncurated data is feasible and effective.

Abstract

Recently, self-supervised learning methods like MoCo, SimCLR, BYOL and SwAV have reduced the gap with supervised methods. These results have been achieved in a control environment, that is the highly curated ImageNet dataset. However, the premise of self-supervised learning is that it can learn from any random image and from any unbounded dataset. In this work, we explore if self-supervision lives to its expectation by training large models on random, uncurated images with no supervision. Our final SElf-supERvised (SEER) model, a RegNetY with 1.3B parameters trained on 1B random images with 512 GPUs achieves 84.2% top-1 accuracy, surpassing the best self-supervised pretrained model by 1% and confirming that self-supervised learning works in a real world setting. Interestingly, we also observe that self-supervised models are good few-shot learners achieving 77.9% top-1 with access to only 10% of ImageNet. Code: https://github.com/facebookresearch/vissl