PaddleOCR-VL now in llama.cpp

The llama.cpp team has integrated the PaddleOCR-VL model into its popular inference engine. This 0.9-billion-parameter model is noted as a top-performing open-source solution for multilingual Optical Character Recognition (OCR). Users can now run efficient, local OCR tasks directly within llama.cpp using provided GGUF file formats, enabling text extraction from images without taxing standard hardware.