Baidu open-sources Unlimited OCR 2 with SR-SWA for unlimited-length document parsing
Baidu_Inc · x · 2026-10-27
Baidu officially released and open-sourced Unlimited OCR 2, targeting the "OCR-3.0 era" of unlimited-horizon document parsing:
- New SR-SWA mechanism condenses text-visual attention interactions into a fixed-size set of working references, avoiding linear attention growth with document length
- Parses virtually any-length documents within a finite context window, with near-constant per-step decoding latency
- MIT-licensed, already at 28.6k GitHub stars and 3k forks
- Supports vLLM inference and ms-swift training; available on Baidu Cloud and a Hugging Face Spaces demo
Related event: Baidu Open-Sources Unlimited OCR 2 for Thousand-Page Document Parsing(5 posts)→
More from Multimodal
- AMBIT: Unified FOA Spatial Audio Generation and Editing from HKUST and Noiz AI — NoizAI · 2026-10-27
- Ornate silver Pioneer CDJ-3000X concept art made with Krea — TitusTeatus · 2026-10-27
- Redditor shows off Ultimate Digital Painting Moss LoRA with striking Cthulhu wallpaper art — kingroka · 2026-10-27
- Midjourney v8.2 style test: one --sref code nails moody manga illustration — azed_ai · 2026-10-27
- A camera movement prompt that makes every second of AI video feel dynamic — umesh_ai · 2026-10-27
- Runware puts FFmpeg in the cloud: common commands via API from $0.02/GB-second — aziz4ai · 2026-10-27