This repo contains the code for "VLM2Vec / MMEB" [ICLR 2025], "VLM2Vec-V2 / MMEB-V2" [TMLR 2026], and "MMEB-V3" [COLM 2026]
DR Pixel: Page-Native Deep Research with Vision-Language Models — benchmark, evaluation harness and training code
Jina VDR is a multilingual, multi-domain benchmark for visual document retrieval
Embedding model prioritized towards Multimodal RAG, overall + VisDoc double top1 on MMEB benchmark
Official ICML 2026 implementation. A framework for training highly efficient list-wise multimodal rerankers for long documents.
Efficient visual document retrieval: NanoVDR, DistilVDR and ColNanoVDR. Small query and document encoders distilled from frozen VLM retrievers.
Benchmarking search agents for closed-corpus, multi-hop retrieval over visually rich documents.
MAGIC: Marginal-Guided Compression with Optimal Transport for Efficient Visual Document Retrieval
Curated papers on multimodal retrieval, multimodal RAG, visual document retrieval, and agentic multimodal search
Measuring presentation sensitivity and style bias in OCR-free visual document retrieval: identical text rendered in controlled styles, evaluated across open retrievers.
Multimodal trade-document clearance copilot — visual document retrieval, grounded extraction with pixel citations, and cross-document reconciliation.
To associate your repository with the visual-document-retrieval topic, visit your repo's landing page and select "manage topics."