Semi-Supervised Vision Transformers
This addresses the challenge of improving semi-supervised learning for image classification, particularly for scenarios with scarce labeled data, representing a strong specific gain rather than a foundational advancement.
The paper tackles the problem of Vision Transformers performing poorly in semi-supervised image classification with limited labeled data, introducing Semiformer, a joint framework that combines transformer and convolutional streams with a fusion module, achieving 75.5% top-1 accuracy on ImageNet and outperforming state-of-the-art methods.
We study the training of Vision Transformers for semi-supervised image classification. Transformers have recently demonstrated impressive performance on a multitude of supervised learning tasks. Surprisingly, we show Vision Transformers perform significantly worse than Convolutional Neural Networks when only a small set of labeled data is available. Inspired by this observation, we introduce a joint semi-supervised learning framework, Semiformer, which contains a transformer stream, a convolutional stream and a carefully designed fusion module for knowledge sharing between these streams. The convolutional stream is trained on limited labeled data and further used to generate pseudo labels to supervise the training of the transformer stream on unlabeled data. Extensive experiments on ImageNet demonstrate that Semiformer achieves 75.5% top-1 accuracy, outperforming the state-of-the-art by a clear margin. In addition, we show, among other things, Semiformer is a general framework that is compatible with most modern transformer and convolutional neural architectures. Code is available at https://github.com/wengzejia1/Semiformer.