Unlimited-OCR Unlimited-OCR is Baidu’s open-source OCR model for one-shot long-horizon document parsing, designed to read multi-page documents in a single pass instead of chunking them. github.com/baidu/Unlimite…

@PythonHub Nice release! Baidu’s Unlimited-OCR uses Reference Sliding Window Attention on top of DeepSeek-OCR to parse entire multi-page documents in one shot with constant memory. Great for long PDFs without chunking. MIT licensed and open-source — worth testing.