Upon launching, users can select from options like Model Fine-tuning, RAG Project, Phi-2 Model Playground, and Windows Optimized Models. The Model Fine-tuning option allows users to initiate a local fine-tuning session using QLoRA, select a model from the catalog, and adjust various settings. The Windows Optimized Models option provides a collection of publicly available AI models optimized for Windows. The studio also welcomes contributions and suggestions from users.
Key takeaways:
- Windows AI Studio is a tool that simplifies generative AI app development by integrating AI development tools and models from Azure AI Studio Catalog and other catalogs like Hugging Face.
- The tool is currently in preview and only runs on NVIDIA GPUs. It is shipped as a Visual Studio Code Extension and requires WSL Ubuntu distro 18.4 or greater.
- Windows AI Studio allows users to browse the AI models catalog, download them locally, fine-tune, test and use them in Windows applications. It also plans to integrate ORT/DML for running AI models on any Windows Hardware.
- The tool also provides a collection of publicly available AI models already optimized for Windows, stored in different locations including Hugging Face, GitHub and others.