PaddlePaddle Releases Open-Source PP-OCRv6 OCR Series With 50 Languages, Up to 5.2x Faster CPU Inference
PaddlePaddle released PP-OCRv6, an Apache 2.0 open-source OCR model series with tiny, small and medium variants ranging from 1.5 million to 34.5 million parameters. The company said the new models improve detection accuracy by 4.9% and recognition accuracy by 5.1% versus PP-OCRv5, while delivering up to 5.2 times faster CPU inference with OpenVINO and support for 50 languages in one unified model.
The release is aimed at broader deployment across browsers, edge devices and servers. PaddlePaddle said PP-OCRv6 adds support for scenarios including PCB, CAD drawings, digital tubes and dot-matrix text, while also covering handwritten, printed, industrial, screen and card text.
From the sources (4 posts)
@vanstriendanielRT @PaddlePaddle: 🚀PP-OCRv6 is officially released！ 🔥PaddleOCR’s new OCR model series scales from 1.5M to 34.5M parameters, bringing stron…
@vanstriendanielRT @AdinaYakup: PP-OCRv6 just released by Baidu @PaddlePaddle ✨ tiny 1.5M / small 7.7M / medium 34.5M ✨ 48+ languages ✨ Supports handwri…
@paddlepaddle🚀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 ne
@adinayakupPP-OCRv6 just released by Baidu @PaddlePaddle ✨ tiny 1.5M / small 7.7M / medium 34.5M ✨ 48+ languages ✨ Supports handwritten/printed/industrial/screen and card text ✨ Edge friendly deployment