Suzhou曙光 and Hunan Hui Shiwei Intelligent Technology recently jointly launched the "Juzhou" V1.5 edge version and its application scenarios in Changsha. This is an important upgrade version of the first pre-trained visual foundation large model in China, which is based on domestic computing power training, marking a new breakthrough in the field of domestic AI visual technology for edge deployment.
Compared to the first version, "Juzhou" V1.5 has completed a strategic transformation from "single system" to "full ecosystem". The new version fully supports the Android system and builds a cross-platform, multi-chip edge AI deployment capability. This technical architecture allows the model to run flexibly on different devices and system environments, removing obstacles for its large-scale application in multiple industry scenarios such as industry, security, and healthcare.
Image source note: The image was generated by AI, and the image licensing service provider is Midjourney.
In terms of technological research and development, the domestic computing power platform provided by Suzhou曙光 has become the core support for the rapid iteration of "Juzhou". During the collaboration, both technical teams have solved a series of technical challenges, including the design of the model base, optimization of the training network, and model feature distillation, verifying the maturity of domestic computing power in the field of large model training, and promoting the coordinated development of domestic hardware and software ecosystems.
This launch is not only a key progress for Hui Shiwei in the field of AI vision, but also reflects the continuous investment of Changsha in the artificial intelligence industry. As a representative project of local AI industry development, the implementation of the "Juzhou" large model provides practical technical support and application demonstration for Changsha to build a "City of AI Applications".
From an industrial perspective, the release of "Juzhou" V1.5 confirms the feasibility of the entire domestic AI technology stack, from the underlying computing power to the upper-level applications, adding a new example for domestic AI vision technology to participate in market competition.
