PaddleOCR 3.5: Running OCR and Document Parsing Tasks with a Transformers Backend

PaddleOCR 3.5 has been released, integrating a Transformers backend to enhance OCR and document parsing capabilities. This version introduces support for various transformer-based models, improving accuracy and performance on benchmark tasks such as the ICDAR 2019 and COCO-Text datasets. The update is significant for practitioners as it allows for more robust text recognition and layout analysis, leveraging state-of-the-art transformer architectures in their applications.