2024/04/30 by Weiquan Huang, Huang, Weiquan, Shen Yifei +3
Computer Science · #Advanced Neural Network Applications #Computer Vision and Pattern Recognition (cs.CV) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis
paper · pdf · doi:10.48550/arxiv.2404.19394
openalex publication_date 2024/04/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
State space models and Mamba-based models have been increasingly applied across various domains, achieving state-of-the-art performance. This technical report introduces the first attempt to train a transferable Mamba model utilizing contrastive language-image pretraining (CLIP). We have trained Mamba models of varying sizes and undertaken comprehensive evaluations of these models on 26 zero-shot classification datasets and 16 out-of-distribution (OOD) datasets. Our findings reveal that a Mamba model with 67 million parameters is on par with a 307 million-parameter Vision Transformer (ViT) model in zero-shot classification tasks, highlighting the parameter efficiency of Mamba models. In tests of OOD generalization, Mamba-based models exhibit exceptional performance in conditions of OOD image contrast or when subjected to high-pass filtering. However, a Hessian analysis indicates that Mamba models feature a sharper and more non-convex landscape compared to ViT-based models, making them more challenging to train. The source code is available at https://github.com/raytrun/mamba-clip.