PaddleOCR-VL-0.9B is a vision-language model developed based on the PaddlePaddle framework, specifically designed for the task of converting image text to text. This model is a reproduction of the official PaddlePaddle version and supports the extraction and recognition of text content from images.

MultimodalTransformers

The latest generation of English text line recognition model developed by the PaddleOCR team, designed for efficient and accurate English OCR recognition, with excellent performance on mobile devices.

PP-DocLayout-L is a high-precision document layout region localization model based on the RT-DETR-L architecture, supporting the detection of 23 common document layout categories.

A dedicated text line recognition model for Devanagari in the PP-OCRv3_rec series developed by the PaddleOCR team, supporting Devanagari recognition with an average accuracy of 96.44%.

PP-LCNet_x1_0_table_cls is an efficient table classification model used to classify input table images and supports the classification of lined tables and borderless tables.

RT-DETR-L_wireless_table_cell_det is a high-precision table cell detection model designed specifically for table recognition tasks. It can accurately locate and mark each cell area in the table image.

UVDoc is mainly used to perform geometric transformations on text images to correct problems such as distortion, tilt, and perspective distortion of documents in the images, thereby improving the accuracy of subsequent text recognition.

PP-OCRv4_server_rec is a text line recognition model in the PP-OCRv4_rec series developed by the PaddleOCR team. It supports text line recognition in general Chinese and English scenarios, mainly focusing on Chinese.

An ultra-lightweight English text line recognition model developed by the PaddleOCR team, supporting the recognition of English and numeric characters

SLANeXt_wired is a deep learning model for table structure recognition, which can convert non - editable table images into editable table formats (such as HTML).

SLANet is a model for table structure recognition that can convert non-editable table images into editable table formats (such as HTML).

RT-DETR-L_wired_table_cell_det is a key module in the table recognition task, mainly responsible for locating and marking each cell area in the table image.

An ultra-lightweight Korean text line recognition model that supports the recognition of Korean and numeric characters, with an average accuracy of 60.21%.

PP-DocBlockLayout is a document layout block positioning model trained based on RT-DETR-L, which can effectively identify layout regions in various document types.

SLANet_plus is a model for table structure recognition that can convert non-editable table images into editable table formats (such as HTML). It plays an important role in the table recognition system and can effectively improve the accuracy and efficiency of table recognition.

PP-OCRv3_mobile_rec is a lightweight text line recognition model developed by the PaddleOCR team. It uses the SVTR algorithm and supports Chinese and English recognition, especially focusing on Chinese scenarios.

An ultra-lightweight Japanese text line recognition model developed by the PaddleOCR team, supporting the recognition of Japanese and numeric characters.

A document text recognition model enhanced based on PP-OCRv4_server_rec, supporting over 15,000 characters, including traditional Chinese characters, Japanese characters, and special symbols.

PP-FormulaNet_plus-M is an enhanced formula recognition model developed by the PaddleOCR team. It supports Chinese formula recognition and improves the processing ability for complex formulas.

An efficient mobile seal text detection model optimized for terminal devices