Data Engine

This project is produce massive MLLM data using data engine. It includes:

• Document: arvix source data, with png images and according text;
• ScreenShots: detected texts along with screenshot;
• Persons: Identify famous persons as world knowledge;

test with latex:

$a^2 = b^2 + c^2$

第一部

全部采用全新的安卓搜索疫情。