ModernVBERT: Towards Smaller Visual Document Retrievers

By Paul Teiletche, Quentin Macé, Max Conti, Ant'onio Loison, Gautier Viaud, Pierre Colombo, Manuel Faysse. 2025

Subfields: Information Retrieval & Search, Multimodal Learning, Efficiency & Compression

Citations in the corpus

• Fantastic (small) Retrievers and How to Train Them: mxbai-edge-colbert-v0 Tech Report (2025, unknown): “ColBERT models have been noted for their particularly robust out-of-domain performance [25], especially in multi-modal settings [27].”
• Unlocking Multimodal Document Intelligence: From Current Triumphs to Future Frontiers of Visual Document Retrieval (2026, unknown): “Addressing efficiency, ColModernVBERT (Teiletche et al., 2025) demonstrates that a compact, bidirectional language encoder can be effectively aligned with a vision encoder to generate granular embeddings; while MetaEmbed (Xiao et al., 2025d) introduces a fixed set of learnable "Meta Tokens" whose fi”
• Evo-Retriever: LLM-Guided Curriculum Evolution with Viewpoint-Pathway Collaboration for Multimodal Document Retrieval (2026, unknown): “However, the static strategy continues to provide these now-unchallenging samples, which ultimately hinders the model’s discriminative ability [10, 13, 17, 34].”