awesome-semi-supervised-learning

Why we included this project

Labeled data is expensive to produce, and semi-supervised learning exists to make the most of the unlabeled data you already have. This list maps that field in a practical way: papers are grouped by task, so image classification, segmentation, object detection, NLP, graph methods, generative models, and theory each get their own section, and you can go straight to what matches your problem. Most entries link to the paper and, when it exists, the code, which matters when you want to reproduce a method like consistency regularization or self-training rather than just read about it. The list also collects ready-to-run codebases such as TorchSSL and Microsoft's unified SSL benchmark, so you have somewhere concrete to start experiments. If your team is weighing which semi-supervised approach to try, this is a good first stop for surveying the options.