Model Releases

PP-OCRv6 on Hugging Face: 50-language OCR with 1.5M to 34.5M parameters

Original title:PP-OCRv6 on Hugging Face: 50-Language OCR from 1.5M to 34.5M Parameters

AI summary

PP-OCRv6 is the latest generation of PaddleOCR's general OCR model family, offering three tiers: tiny (1.5M), small (7.7M) and medium (34.5M). Medium and small support 50 languages (Simplified and Traditional Chinese, English, Japanese and 46 Latin-script languages). On official multi-scenario benchmarks, medium achieves 86.2% detection Hmean and 83.2% recognition accuracy, up 4.6 and 5.1 percentage points respectively over PP-OCRv5_server. The models use a unified PPLCNetV4 backbone, RepLKFPN detection module and EncoderWithLightSVTR recognition module, and can be flexibly deployed via backends such as PaddleOCR, Transformers and ONNX Runtime.

Key Takeaways

  • PP-OCRv6 is the latest generation of PaddleOCR's general OCR model family,offering three tiers:…
  • 5M),small (7.
  • 7M) and medium (34.
  • Medium and small support 50 languages (Simplified and Traditional Chinese,English,Japanese and 46…
  • On official multi-scenario benchmarks,medium achieves 86.

Quick Facts

Published
Updated
Source
Hugging Face 社区博客(混合发现)
Category
Model Releases
AI score
69
Read the original Read the full article at the original source
Why it matters

OCR模型并没有因为VLM的兴起而消失,PP-OCRv6用1.5M到34.5M参数覆盖50种语言,对需要轻量、准确OCR的产品来说,这是一个务实的选择。

Related entities

Source & copyright

This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.

Original article · AIHOT entry

How to Cite

When citing this article, keep the title, date, and permalink, and credit Orange News as the source.

Orange News: PP-OCRv6 on Hugging Face: 50-language OCR with 1.5M to 34.5M parameters. June 22, 2026. https://new.nat6.net/en/news/cmqpamzm108huslx6vf1oonz9/

View all →