Learning To Classify Images Without Labels
A two-step approach where feature learning and clustering are decoupled.
self-supervised-learning unsupervised-learning image-classification clustering computer-vision tutorial research code paper arxiv:2005.12320

Note: code + pretrained models + configuration files will be released (in a few weeks) to produce numbers even better than in the current version of the paper.

Don't forget to tag @wvangansbeke in your comment, otherwise they may not be notified.

Authors community post
PhD researcher at KU Leuven. Especially interested in computer vision, machine learning and deep learning.
Share this project
Similar projects
SwAV
Unsupervised Learning of Visual Features by Contrasting Cluster Assignments
SimCLR with PyTorch Lightning
A lightweight implementation of SimCLR with PyTorch Lightning. Will be used as the start point for modifications on top of SimCLR to experiment.
SimCLR - Contrastive Learning of Visual Representations
How to load pretrained/finetuned SimCLR models from hub modules for fine-tuning.
CS294-158-SP19 Deep Unsupervised Learning
This course will cover two areas of deep learning in which labeled data is not required: Deep Generative Models and Self-supervised Learning.
Top collections