With the rapid development of artificial intelligence technology, large models have become an important field. However, running large models requires huge amounts of computing power and high costs, which is prohibitive for many individuals and small businesses. However, with the release of the Intel Core Ultra series of processors, this is changing dramatically.
At the Core Ultra launch event, the Codeshell multilingual** model was moved into a thin and light laptop. This means that the future operation of large models does not necessarily depend only on servers or desktops, and the thin and light notebooks that you carry with you can also become the main battlefield for the operation of large models. The codeshell model can be seen as an AI assistant that is good at writing, checking for defects, and completing repairs and tests. With the help of this large model, the development team can detach the focus of development from heavy writing and bug fixing, and concentrate resources on more important functional planning, creative implementation and other work to promote the software development process.
Codeshell is quite famous in the industry, it was developed by the Knowledge Computing Laboratory of Peking University, as of November this year, HUGGING FACE international authoritative open source ** large model list, Codeshell ranked first in the same scale of multi-language ** large model. At the same time, the CodeShell team has also open-sourced a series of intelligent assistant solutions, including open-source VS Code and JetBrains IDE plug-ins, covering common software development activities such as generation, defect detection and repair, and test case generation.
In view of the problems that the model training evaluation benchmark cannot accurately reflect its real capabilities and the evaluation scenario is disconnected from the actual development environment, the Codeshell team proposes a multi-task programming model evaluation benchmark with a complete program context, which provides complete program context information, covering common software engineering tasks including ** generation, test case generation, defect detection, defect repair, etc.
What's more, the Codeshell model is already capable of being deployed and run on thin and light laptops. In the future, with the continuous optimization of the AI acceleration of Intel Core Ultra series processors, the performance of CodeShell large models on thin and light notebooks will be further improved, and it is expected to become the main platform for large models to run in the future.
The Intel Core Ultra series processors are the first processors built on Intel 4 process technology and represent the biggest evolution of Intel architecture in 40 years. It uses advanced FoverOS 3D packaging technology, combining cutting-edge intellectual property achievements with advanced process technology to optimize the performance of related modules.
In the future, a large number of professional applications, including design, content creation, and engineering planning, will also bloom on AI PCs. It is expected that by 2025, more than 100 million AI PCs will be shipped worldwide. The cooperation involves 100 ISV partners, covering more than 300 AI acceleration functions, and AI applications will be promoted more widely. As the proposer and important promoter of the concept of AI PC, Intel will continue to invest more resources to make AI and PC highly integrated at the ecological level, so that AI PC will become a powerful tool for every developer and creator.