Tongji University needed a domestic AI stack with stronger computing and model-tooling support for embodied intelligence, multimodal models, autonomous perception and multi-agent research.
Reported impact
Tongji gained a domestic full-stack research platform that speeds model tuning, supports embodied-intelligence applications and reinforces teaching programs in intelligent software and urban sensing.
Tongji University and Baidu AI Cloud built a domestically based embodied-intelligence large-model development platform that combines Kunlun P800 accelerators, the BaiGe heterogeneous-compute platform, the Qianfan model platforms and the Yijian vision-model platform. Rather than using these components separately, Tongji assembled them into a one-stop research environment for tuning, adapting and evaluating large models in embodied-intelligence, autonomous sensing and software-engineering scenarios. The system supports deep adaptation of ERNIE and DeepSeek series models, then exposes model-effect tuning toolchains and research application capabilities to university projects. That lets Tongji use one governed infrastructure base for research and teaching in robots, drones, smart-city sensing, multi-agent coordination and embedded software. The platform is therefore both an academic compute substrate and a reusable model-development environment: it helps the university manage specialized training and inference work while also supplying a common stack for talent development and cross-domain software education.
Research data and model experiments are scheduled across Kunlun and BaiGe resources, tuned through the Qianfan toolchain, and then surfaced into embodied-intelligence applications for robots, drones and smart-city sensing research.
The platform supports robot, drone, multi-agent and smart-city teaching and research scenarios (original: 全面服务于物联网与智慧城市、智慧城市与智能运维、社区典型场景智慧服务一体化、机器自主感知、多智能体与具身智能、智能驾驶与嵌入式软件等特色化软件人才培养).
Tongji and Baidu deeply adapted ERNIE and DeepSeek series models and formed both a large-model R&D platform and a research-application platform (original: 同济大学与百度智能云基于昆仑芯P800与文心大模型及DeepSeek系列模型深度适配，并基于百度智能云千帆大模型平台，形成了具身智能大模型研发平台和具身智能科研应用平台).
Tongji selected Kunlun P800, BaiGe, Qianfan and Yijian to build a fully domestic embodied-intelligence large-model development platform (original: 项目方案选用性能卓越的昆仑芯P800、百度百舸AI异构计算平台、百度智能云千帆大模型平台、百度智能云一见视觉大模型平台进行融合).
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Baidu AI Cloud Sample Oct 13, 2025 Kaoshibao (Shanghai Juxian Network Technology Co., Ltd.) Kaoshibao needed to use new AI interaction and content technologies to improve personalized learning, lower content-production costs, and make very large question banks easier to ingest, search, and explain. "Zero-basis" users improved certificate pass rates 35%. Intelligent Search and Generation Component · Multimodal Recognition · Qianfan AppBuilder Baidu AI Cloud Sample May 15, 2025 Lanxum Education (立思辰) Lanxum wanted to make Chinese-language learning more engaging and personalized, but it needed a structured way to connect teaching content, teacher knowledge, dialogue, and student guidance instead of relying on static… Chinese Teaching Research Knowledge Graph · Intelligent Voice Dialogue System · Virtual Teacher Baidu AI Cloud Sample May 15, 2025 Shanghai Jiao Tong University Shanghai Jiao Tong University wanted to connect frontier AI techniques with real research workflows so faculty could use scientific data, models and computing resources more efficiently instead of relying on years of… A molecular SMILES query can return reaction processes, conditions and literature links in about 5 seconds. AI Middle Platform · BaiGe Heterogeneous Compute Platform · Data Middle Platform 