SILC: Improving Vision Language Pretraining with Self-distillation


METADATA ONLY

Date

2025

Publication Type

Conference Paper

ETH Bibliography

yes

Citations

Altmetric
METADATA ONLY

Data

Rights / License

Abstract

Image-Text pretraining on web-scale image caption datasets has become the default recipe for open vocabulary classification and retrieval models thanks to the success of CLIP and its variants. Several works have also used CLIP features for dense prediction tasks and have shown the emergence of open-set abilities. However, the contrastive objective used by these models only focuses on image-text alignment and does not incentivise image feature learning for dense prediction tasks. In this work, we introduce SILC, a novel framework for vision language pretraining. SILC improves image-text contrastive learning with the simple addition of local-to-global correspondence learning by self-distillation. We show that distilling local image features from an exponential moving average (EMA) teacher model significantly improves model performance on dense predictions tasks like detection and segmentation, while also providing improvements on image-level tasks such as classification and retrieval. SILC models sets a new state of the art for zero-shot classification, few shot classification, image and text retrieval, zero-shot segmentation, and open vocabulary segmentation. We further show that SILC features greatly benefit open vocabulary detection, captioning and visual question answering.

Permanent link

Publication status

published

Book title

Computer Vision – ECCV 2024

Volume

15079

Pages / Article No.

38 - 55

Publisher

Springer

Event

18th European Conference on Computer Vision (ECCV 2024)

Edition / version

Methods

Software

Geographic location

Date collected

Date created

Subject

Vision Language Models; Self-supervised Learning; Open-Vocabulary

Organisational unit

03514 - Van Gool, Luc (emeritus) / Van Gool, Luc (emeritus) check_circle

Notes

Funding