Add Nanonets-OCR2-3B and HunyuanOCR models, fix 90-degree flowchart rotation and lane parsing

Upgrade all 7 model adapters and pipelines matching official Hugging Face demo architectures

Upgrade OCR Space with 2-Column Layout Engine, Document AI Reasoning, Multi-Page PDF pagination, and accurate latency timing

Deploy self-contained 7-Model OCR Benchmark with White Theme to Hugging Face Space

Configure Docker SDK on Hugging Face Space with FastAPI and React White Theme

Redesign UI/UX Pro Max with Light Theme and memory-safe PDF pipeline

Add editable backend URL configuration and connection helper

Deploy 7-Model OCR Benchmark Space with Bounding Box Visualizer