PaliGemma: A versatile 3B VLM for transfer

作者：Lucas Beyer, Andreas Steiner, André Susano Pinto, Alexander Kolesnikov, Xiao Wang, Daniel Salz, Maxim Neumann, Ibrahim Alabdulmohsin, Michael Tschannen, Emanuele Bugliarello, Thomas Unterthiner, Daniel Keysers, Skanda Koppula, Fangyu Liu, Adam Grycner, Alexey A. Gritsenko, Neil Houlsby, Manoj Kumar, Keran Rong, Julian Martin Eisenschlos, Rishabh Kabra, Matthias Bauer, Matko Bošnjak, Xizhang Chen, Matthias Minderer, Paul Voigtlaender, Ioana Bica, Ivana Balažević, Joan Puigcerver, Pinelopi Papalampidi, Olivier J. Hénaff, Xi Xiong, Radu Soricut, Jeremiah Harmsen, Xiaohua Zhai · 发表于：arXiv (Cornell University) · 年份：2024 · DOI：10.48550/arxiv.2407.07726 · 被引用次数：13 · 研究领域：Advanced Fiber Optic Sensors

PaliGemma is an open Vision-Language Model (VLM) that is based on the SigLIP-So400m vision encoder and the Gemma-2B language model. It is trained to be a versatile and broadly knowledgeable base model that is effective to transfer. It achieves strong performance on a wide variety of open-world tasks. We evaluate PaliGemma on almost 40 diverse tasks including standard VLM benchmarks, but also more specialized tasks such as remote-sensing and segmentation.