2 papers
cs.CV2026
Elastic ViTs from Pretrained Models without Retraining
Walter Simoncini, Michael Dorkenwald, Tijmen Blankevoort +2
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 con…
cs.CV2024
No Train, all Gain: Self-Supervised Gradients Improve Deep Frozen Representations
Walter Simoncini, Spyros Gidaris, Andrei Bursuc +1
This paper introduces FUNGI, Features from UNsupervised GradIents, a method to enhance the features of transformer encoders by leveraging self-supervised gradients. Our method is s…