Ai2 发布 olmOCR 2，可将文档图像转为可检索文本并开放权重与代码
olmOCR 2 turns document images into searchable text. Its open weights & code let researchers run it locally or adapt it to their libraries.
摘要
Ai2 发布 olmOCR 2，可将文档图像转换为可检索文本，并开放模型权重与代码，研究者可在本地运行或针对自有馆藏做适配。推文提到两个历史与考古研究项目，用来说明这一能力可以支持哪些工作。
摘要由模型根据原文撰写，以原文为准。
正文 · 原文
olmOCR 2 turns document images into searchable text. Its open weights & code let researchers run it locally or adapt it to their libraries.
Two projects show what that makes possible for historical & archaeological research. 🧵
来源：Ai2 · x.com
引用原文
allen_ai. olmOCR 2 turns document images into searchable text. Its open weights & code let researchers run it locally or adapt it to their libraries.[EB/OL]. (2026-10-29)[访问日期]. https://x.com/allen_ai/status/2115511924091724039.