Microsoft recently launched an AI PC device for Windows 11 operating system.Open source development application resources called "AI Dev Gallery", allowing developers to import device-side artificial intelligence application functions into application services.
The "AI Dev Gallery" open-source development application resources allow developers to use various AI models and import various AI application functions according to application design requirements. At the same time, it is built for device-side execution requirements, so it can be used without an Internet connection. The execution speed is also faster than large-scale AI models that require an Internet connection.
Currently, applications running on Windows 10 and Windows 11, as well as x64 and Arm64 architectures, are compatible with the "AI Dev Gallery" open-source development resources. To utilize this support, devices must have at least 16GB of memory and at least 20GB of storage. For faster, more computationally intensive applications, a dedicated graphics card with 8GB of built-in video memory is recommended.
Although more and more companies are launching various large-scale natural language models (LLMs) to promote more natural artificial intelligence interactive user experiences, since most of them rely on network connections, their execution efficiency is often affected by the network environment. Currently, more companies have begun to develop small-scale natural language models (SLMs) and improve artificial intelligence models through fine-tuning and distillation. Their performance in specific execution can approach that of large natural language models, while also maintaining a higher efficiency execution speed.
In addition to its ongoing collaboration with OpenAI, Microsoft has also proposedPhi series small natural language models, boasting to be able to execute with lighter features, and plansApplication within its service, thereby reducing the execution cost of artificial intelligence application services.
