3 accepted papers
Vision foundation models achieve remarkable performance but are only available in a limited set of pre-determined sizes, forcing sub-optimal deployment choices under real-world constraints. We introduce SnapViT: single-shot network approximation for pruned Vision Transformers, a new post-pretraining…
This paper introduces FUNGI, **F**eatures from **UN**supervised **G**rad**I**ents, a method to enhance the features of transformer encoders by leveraging self-supervised gradients. Our method is simple: given any pretrained model, we first compute gradients from various self-supervised objectives fo…
Named Entity Recognition is a fundamental task in information extraction and is an essential element for various Natural Language Processing pipelines. Adversarial attacks have been shown to greatly affect the performance of text classification systems but knowledge about their effectiveness against…