OCR

The 'OCR' dataset/benchmark contains tasks related to Optical Character Recognition and is used to evaluate the performance of Multimodal Large Language Models (MLLMs) in recognizing and interpreting text within images.

Papers using OCR (5)

LoMo: Local Modality Substitution for Deeper Vision-Language Fusion2026Text-Guided Layer Fusion Mitigates Hallucination in Multimodal LLMs2026Inverse-LLaVA: Rethinking Multimodal Alignment via Text-to-Vision Mapping2025Investigating Redundancy In Multimodal Large Language Models With Multiple Vision Encoders2025Optmerge: Unifying Multimodal LLM Capabilities And Modalities Via Model Merging2025