GitHub

Skip Tuning: Pre-trained Vision-Language Models are Effective and Efficient Adapters Themselves

Offical implementation of the paper Skip Tuning: Pre-trained Vision-Language Models are Effective and Efficient Adapters Themselves. Abstract Prompt tuning (PT) has long been recognized as an ...
IEEE

Forensics Adapter: Adapting CLIP for Generalizable Face Forgery Detection

Abstract: We describe the Forensics Adapter, an adapter network designed to transform CLIP into an effective and generalizable face forgery detector. Although CLIP is highly versatile, adapting it for ...
IEEE

Tip-Adapter-SA: Training-Free CLIP-Adapter Enhanced by Semantic Alignment for Multi-Label Few-Shot Classification with Single Positive Labels

Abstract: Multi-label classification is a fundamental task that requires predicting all applicable labels for each sample. Previous methods often rely heavily on training models with large-scale multi ...