The PP-OCRv6 model family offers a range of sizes from 1.5M to 34.5M parameters. It supports 50 languages with a focus on high-speed inference and accuracy. This release integrates the toolkit into the Hugging Face ecosystem. Developers can now deploy lightweight, multilingual text recognition without managing complex custom environments.