Model Recipes
PaddlePaddle

PaddlePaddle/PaddleOCR-VL

PaddleOCR-VL (0.9B) — compact vision-language model for document parsing, OCR, tables, formulas, charts

View on HuggingFaceView on ModelScope
dense0.9B131,072 ctxvLLM 0.11.1+multimodal
Guide