We introduce BLIP3-KALE, a dataset of 218 million image-text pairs that bridges the gap between descriptive synthetic captions and factual web-scale alt-text. KALE augments synthetic dense image captions with web-scale alt-text to generate factually grounded image captions. Our two-stage approach leverages large vision-language models and language models to create knowledge-augmented captions, which are then used to train a specialized VLM for scaling up the dataset. We train vision-language models on KALE and demonstrate improvements on vision-language tasks. Our experiments show the utility of KALE for training more capable and knowledgeable multimodal models. We release the KALE dataset at https://huggingface.co/datasets/Salesforce/blip3-kale
@article{arxiv.2411.07461,
title = {BLIP3-KALE: Knowledge Augmented Large-Scale Dense Captions},
author = {Anas Awadalla and Le Xue and Manli Shu and An Yan and Jun Wang and Senthil Purushwalkam and Sheng Shen and Hannah Lee and Oscar Lo and Jae Sung Park and Etash Guha and Silvio Savarese and Ludwig Schmidt and Yejin Choi and Caiming Xiong and Ran Xu},
journal= {arXiv preprint arXiv:2411.07461},
year = {2024}
}