Xray-Visual Models: Scaling Vision models on Industry Scale Data

TL;DR

Xray-Visual is a large-scale, multimodal vision model trained on industry data, achieving state-of-the-art results and robustness across various benchmarks by combining advanced training strategies and efficient architecture.

Contribution

The paper introduces Xray-Visual, a novel scalable vision model architecture trained on billions of social media image-text and video-hashtag pairs, with a unique three-stage training pipeline and enhanced efficiency.

Findings

Achieves state-of-the-art performance on ImageNet, Kinetics, HMDB51, and MSCOCO.

Demonstrates robustness to domain shifts and adversarial attacks.

Integrating large language models improves retrieval and generalization.

Abstract

We present Xray-Visual, a unified vision model architecture for large-scale image and video understanding trained on industry-scale social media data. Our model leverages over 15 billion curated image-text pairs and 10 billion video-hashtag pairs from Facebook and Instagram, employing robust data curation pipelines that incorporate balancing and noise suppression strategies to maximize semantic diversity while minimizing label noise. We introduce a three-stage training pipeline that combines self-supervised MAE, semi-supervised hashtag classification, and CLIP-style contrastive learning to jointly optimize image and video modalities. Our architecture builds on a Vision Transformer backbone enhanced with efficient token reorganization (EViT) for improved computational efficiency. Extensive experiments demonstrate that Xray-Visual achieves state-of-the-art performance across diverse benchmarks, including ImageNet for image classification, Kinetics and HMDB51 for video understanding, and MSCOCO for cross-modal retrieval. The model exhibits strong robustness to domain shift and adversarial perturbations. We further demonstrate that integrating large language models as text encoders (LLM2CLIP) significantly enhances retrieval performance and generalization capabilities, particularly in real-world environments. Xray-Visual establishes new benchmarks for scalable, multimodal vision models, while maintaining superior accuracy and computational efficiency.