Skip to content

ExplainableML/flair

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 

Repository files navigation

FLAIR: VLM with Fine-grained Language-informed Image Representations

Authors: Rui Xiao, Sanghwan Kim, Mariana-Iuliana Georgescu, Zeynep Akata, Stephan Alaniz

Abstract

CLIP has shown impressive results in aligning images and texts at scale. However, its ability to capture detailed visual features remains limited because CLIP matches images and texts at a global level. To address this issue, we propose FLAIR, Fine-grained Language-informed Image Representations, an approach that utilizes long and detailed image descriptions to learn localized image embed dings. By sampling diverse sub-captions that describe fine-grained details about an image, we train our vision- language model to produce not only global embeddings but also text-specific image representations. Our model introduces text-conditioned attention pooling on top of local image tokens to produce fine-grained image representations that excel at retrieving detailed image content. We achieve state-of-the-art performance on both, existing mul- timodal retrieval benchmarks, as well as, our newly introduced fine-grained retrieval task which evaluates vision- language models’ ability to retrieve partial image content. Furthermore, our experiments demonstrate the effectiveness of FLAIR trained on 30M image-text pairs in capturing fine-grained visual information, including zero-shot semantic segmentation, outperforming models trained on billions of pairs.

Methodology

Code and Model

Code and pre-trained models will be released soon.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published