PaddleOCR 3.5: Running OCR and Document Parsing Tasks with a Transformers Backend
• 36
AI & ML interests
Deep Learning Framework
Recent Activity
View all activity
Papers
PaddleOCR-VL-1.6: Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Training
GraphNet: A Large-Scale Computational Graph Dataset for Tensor Compiler Research
Articles
From 1.5M to 34.5M Parameters, Surpassing Billion-Scale VLMs on OCR Tasks
-
PaddlePaddle/PP-OCRv6_medium_det_safetensors
Image-to-Text • 22M • Updated • 13 -
PaddlePaddle/PP-OCRv6_small_det_safetensors
Image-to-Text • 2.48M • Updated • 7 -
PaddlePaddle/PP-OCRv6_tiny_det_safetensors
Image-to-Text • 438k • Updated • 7 -
PaddlePaddle/PP-OCRv6_medium_rec_safetensors
Image-to-Text • 19.2M • Updated • 6
Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsing
PP-StructureV3 is a SOTA document parsing solution on OmniDocBench, supporting the conversion of PDFs and do cument images to Markdown and JSON.
Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Training
Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model
-
PaddlePaddle/PaddleOCR-VL
Image-Text-to-Text • 1.0B • Updated • 7.09k • 1.62k -
PaddleOCR-VL Online Demo
📈251Extract text, tables, formulas, and charts from images
-
PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model
Paper • 2510.14528 • Published • 129 -
PaddlePaddle/PP-DocLayoutV2
Object Detection • Updated • 5.77k • 31
PP-OCRv5 is the latest text recognition solution, supporting Simplified Chinese, Chinese Pinyin, Traditional Chinese, English, and Japanese
-
PaddlePaddle/arabic_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 2.07k • 4 -
PaddlePaddle/chinese_cht_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 57 • 1 -
PaddlePaddle/cyrillic_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 2.03k • 1 -
PaddlePaddle/devanagari_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 1.88k • 1
From 1.5M to 34.5M Parameters, Surpassing Billion-Scale VLMs on OCR Tasks
-
PaddlePaddle/PP-OCRv6_medium_det_safetensors
Image-to-Text • 22M • Updated • 13 -
PaddlePaddle/PP-OCRv6_small_det_safetensors
Image-to-Text • 2.48M • Updated • 7 -
PaddlePaddle/PP-OCRv6_tiny_det_safetensors
Image-to-Text • 438k • Updated • 7 -
PaddlePaddle/PP-OCRv6_medium_rec_safetensors
Image-to-Text • 19.2M • Updated • 6
Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Training
Towards a Multi-Task 0.9B VLM for Robust In-the-Wild Document Parsing
Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model
-
PaddlePaddle/PaddleOCR-VL
Image-Text-to-Text • 1.0B • Updated • 7.09k • 1.62k -
PaddleOCR-VL Online Demo
📈251Extract text, tables, formulas, and charts from images
-
PaddleOCR-VL: Boosting Multilingual Document Parsing via a 0.9B Ultra-Compact Vision-Language Model
Paper • 2510.14528 • Published • 129 -
PaddlePaddle/PP-DocLayoutV2
Object Detection • Updated • 5.77k • 31
PP-StructureV3 is a SOTA document parsing solution on OmniDocBench, supporting the conversion of PDFs and do cument images to Markdown and JSON.
PP-OCRv5 is the latest text recognition solution, supporting Simplified Chinese, Chinese Pinyin, Traditional Chinese, English, and Japanese
-
PaddlePaddle/arabic_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 2.07k • 4 -
PaddlePaddle/chinese_cht_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 57 • 1 -
PaddlePaddle/cyrillic_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 2.03k • 1 -
PaddlePaddle/devanagari_PP-OCRv3_mobile_rec
Image-to-Text • Updated • 1.88k • 1