太初元碁等10余家国产AI芯片深度适配MinerU自研模型

Group 1 - The collaboration between Shanghai AI Laboratory's OpenDataLab team, DeepLink team, and domestic chip manufacturers has successfully adapted over 10 mainstream domestic computing power solutions, including Ascend, PingTouGe, and others, to enhance the ecological compatibility and adaptability of the MinerU project [1] - MinerU's self-developed VLM model achieves an accuracy rate of 99% in capturing elements from PDFs and complex web pages, enabling precise restoration and structured extraction of intricate mathematical formulas and nested structured tables [1] - The core value of MinerU lies in its cross-industry applicability and high parsing accuracy, serving as an efficient data production engine for large model development and a precise document parsing tool for government, enterprise, and research sectors [1] Group 2 - Domestic AI large models have been updated recently, with domestic AI chip companies quickly adapting to these new versions, exemplified by TaiChuang YuanQi, which has completed adaptations for over 30 AI large models, including DeepSeek, QianWen, and others [2] - The adaptations cover a wide range of models, including Qwen3Dense/MoE series, BAAI Embedding/Reranker series, and various multi-modal understanding and generation models, indicating a strong push towards the integration of intelligent computing and industry [2]