← Library · Frontier

Baidu Introduces 'Unlimited OCR' with Constant Memory Use for Multi-Page Documents

Baidu researchers have developed 'Unlimited OCR,' a model capable of processing dozens of document pages in a single pass with constant memory and speed, regardless of text length. This is achieved through a redesigned attention mechanism called Reference Sliding Window Attention (R-SWA). The model, built on Deepseek OCR, uses a 3-billion-parameter Mixture-of-Experts architecture and scores 93.92% on OmniDocBench v1.6.

Why it matters

This innovation addresses a critical bottleneck in long-document processing for OCR, making it far more efficient and scalable for applications requiring the analysis of extensive textual data.

Learn one new AI thing every day.

Daily Deck sends you seven plain-English cards like this every morning. Free.

Start free