Latest news! The Department of Electronic Engineering, Tsinghua University recently has completed the construction of the hardware part of the AI big data platform. The platform is located at the Institute for Electronics and Information Technology in Tianjin, Tsinghua University in China-Singapore Tianjin Eco-city. It has high-performance resource allocations and the physical configuration of some computing nodes exceeds the existing commercial public cloud products.
Relying on the cloud computing resource management system and the HPC cluster management system, the platform can effectively organize the distribution and management of computing resources, and can realize the self-service functions of users. The platform supports mainstream AI algorithm frameworks (Caffe, Caffe2, CNTK, Tensorflow, etc.); Recently, multiple advanced AI computing functions such as face recognition, speech recognition, and emotion analysis on the AI open platform researched and developed by the Research Center for Media Big-data Cognitive Computing of the Department of Electronic Engineering will be transplanted to the platform to provide application services for the society; In the future, research on the six system application of urban intelligent planning system, healthcare management system, driverless car system, AI deep learning training system, video structuralization and analysis system, and Webservice cluster system will be conducted and product incubation attempts will be launched.
The platform supports scientific research and external services simultaneously.
(1)It is connected to the Rohm Building through a dedicated line to provide AI big data computing platform for Tsinghua University and relevant scientific research units;
(2)Provide Internet services to provide support for the services of AI research results for the whole society.
Presently, the purchase and installation and commissioning of all hardware have been completed, and the operation team is being improved. It is expected that the trial run will be started in mid-July.
