PaddlePaddle (@PaddlePaddle) on X (original) (raw)

The first independent R&D and Open-Source deep learning platform in China. Powering the ERNIE model family.

• Pinned
  
  🚀PP-OCRv6 is officially released！ 🔥PaddleOCR’s new OCR model series scales from 1.5M to 34.5M parameters, bringing stronger accuracy, faster inference, and broader deployment options — from browsers and edge devices to servers. 📊What’s new: 🔸Tiny / Small / Medium models:
  
  
• 
  🚀 PaddleOCR-VL is here! Introducing PaddleOCR-VL (0.9B) — the ultra-compact Vision-Language model that reaches SOTA accuracy across text, tables, formulas, charts & handwriting. Breaking the limits of document parsing!🌍 Powered by: • NaViT dynamic vision encoder • ERNIE
  
• 
  PaddleSeg is an easy-to-use image segmentation library with awesome pre-trained model zoo, supporting wide-range of practical tasks in Semantic Segmentation, Interactive Segmentation, Panoptic Segmentation, Image Matting, 3D Segmentation, etc.github.com/PaddlePaddle/P…
  
• 
  🔥 Huge milestone! PaddleOCR-VL just hit #1 on Hugging Face Trending — only 20 hours after release! 🚀 Let’s keep pushing the boundaries of document intelligence — come and drop a ❤️ now! A big thank you to our amazing community for the love & support! 👉
• 
  🚀 PaddleOCR-VL Benchmark Series | Episode 1: Complex Layout Decoding In the fast-evolving world of document intelligence, many models shine — but which truly performs in real-world scenarios? 🌍 We benchmarked PaddleOCR-VL against MinerU2.5, MonkeyOCR, and GPT-4o, comparing
  
  
• 
  💥 Now Open-Sourced: ERNIE-4.5-VL-28B-A3B-Thinking! 🧠 Compact model, powerful multimodal reasoning. 💪 Just 3B activated parameters—yet it rivals flagship giants! ✨ Lightweight scale, near-SOTA performance. Vision, redefined. 🔍 Highlights: • Advanced visual-language
  
  
• 
  🚀Excited to announce that the ERNIE 4.5 series models are officially open-sourced today! 🙌ERNIE 4.5 models achieved state-of-the-art performance across multiple text and multimodal benchmarks, especially in instruction following, world knowledge memorization, visual
  
• 
  🚀 PaddleOCR-VL Benchmark Series | Episode 2: Multilingual Text Recognition In global document workflows, mixed-language texts often lead to confusion and misreads — especially for low-resource languages. 🌏 📊 Key Insight (Ep.2 – Multilingual Recognition): ✅ PaddleOCR-VL •
• 
  🚀 PaddleOCR-VL Benchmark Ep.3 — Handwriting & Vertical Text ✍️ Handwriting & vertical layouts have long been OCR’s toughest challenge. ✅ PaddleOCR-VL • Excels at Chinese & English handwriting — accurate even with messy strokes. • Flawlessly recognizes vertical text in books
• 
  Announcing our open source collaboration with
  @huggingface
  ! 🚀 75+ #PaddleNLP models already on the
  @huggingface
  Hub 🔥 More awesome #PaddlePaddle models across text, image, audio, video and multi-modalities in the works! Learn more at