TL;DR: We propose CUFIT, a robust fine-tuning method for vision foundation models under noisy label conditions, based on the advantages of linear probing and adapters. Download the training data, ...

Abstract: The linear-probing hash table is one of the oldest and most widely used data structures in computer science. However, linear probing famously comes with a major draw-back: as soon as the ...

Abstract: Representing hyperspectral images (HSIs) is a complex and challenging task, primarily due to spectral uncertainty. The learnable prototypical contrastive learning is specialized in ...