Here’s the translation to American English:
—
In a significant step towards democratizing the development of artificial intelligence applications, Microsoft has launched the AI Toolkit for Visual Studio Code (VS Code). This new set of tools is designed to simplify the creation of generative AI applications by integrating advanced models from the Azure AI Foundry catalog and other sources, such as Hugging Face.
With the AI Toolkit, developers can download and run artificial intelligence models locally. This support covers platforms such as Windows 11 with DirectML acceleration, Linux with NVIDIA acceleration, and support for macOS is expected to be available soon. Users have the option to test these models in an intuitive “playground” environment or deploy them in their applications via a REST API.
One of the standout features of the AI Toolkit is its ability to fine-tune AI models, both locally and in the cloud. This functionality allows developers to create new capabilities, improve the reliability of generated responses, and adjust their tone and format. The toolkit also provides guides for optimizing popular language models like Phi-3 and Mistral.
Additionally, it highlights the ability to deploy AI features in the cloud or in functional applications on local devices. Developers can also perform batch runs of inputs for selected models in the playground environment.
For those looking to dive deeper into the AI Toolkit, Dona Sarkar and John Lam have shared a summary in a video available from Microsoft Build 2024. Furthermore, the latest documentation and details on how to download and use this tool can be found on the official Visual Studio Code website, where developers can access tutorials and additional guides.
via: MiMub in Spanish
