PaddleOCR 3.5 發布：支援 Transformers 後端，輕鬆執行 OCR 與文件解析任務
Original: PaddleOCR 3.5: Running OCR and Document Parsing Tasks with a Transformers Backend
The well-known open-source OCR (Optical Character Recognition) toolkit PaddleOCR has long been celebrated for its high accuracy…
知名開源 OCR 工具 PaddleOCR 3.5 正式整合至 Hugging Face Transformers 生態系。開發者現在可以直接使用 PyTorch/Transformers 後端執行文字偵測、識別及表格解析等任務，無需安裝複雜的 PaddlePaddle 框架。這項更新極大地簡化了 OCR 與 RAG（檢索增強生成）等下游任務的整合流程。
The well-known open-source OCR (Optical Character Recognition) toolkit PaddleOCR has long been celebrated for its high accuracy, lightweight models, and strong support for multiple languages — particularly Chinese and English. However, PaddleOCR was previously tightly coupled to Baidu's PaddlePaddle deep learning framework, which often created compatibility challenges for developers accustomed to the PyTorch ecosystem when it came to installation, dependency management, and integration with mainstream Transformer-based models.
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Hugging Face Blog →Summaries are AI-generated; the original article is authoritative.