A Blog post by PaddlePaddle on Hugging Face

HuggingFace's platform is a resource for developers and researchers working in natural language processing (NLP) and machine learning, offering insights into NLP models, tools, and datasets. Through articles, tutorials, and open-source projects, HuggingFace offers insights into state-of-the-art NLP techniques, transformer architectures, and transfer learning methods. Developers can learn about using pre-trained models, fine-tuning strategies, and deploying NLP applications with HuggingFace's libraries and APIs.

Hugging Face

PaddleOCR 3.5 introduces Hugging Face Transformers as a supported inference backend, selectable via the `engine="transformers"` parameter. This allows OCR models like PP-OCRv5 and document parsing models like PaddleOCR-VL 1.5 to run within PyTorch/Transformers-centered stacks without replacing the default Paddle static graph backend. The release targets developers building RAG, Document AI, and agent workflows who want tighter integration with the Hugging Face ecosystem. Backend-specific options such as dtype, device placement, and attention implementation are configurable via `engine_config`. The default Paddle static backend remains recommended for maximum throughput.

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

When should you use the Transformers backend?