In its commitment to Open Source Technology Microsoft has announced a handful of tools on Azure AI to make building more intelligent apps both easier and more accessible. Tools to fine-tune AI deployments highlight the offerings.
Foundation Models in Azure Machine Learning allows users to fine-tune pre-trained open-source models from Hugging Face throughout Azure ML. The partnership with Hugging Face was announced last Spring. Responsible AI Toolbox additions were announced last week. The Responsible Mitigations AI Library and the Fairlearn fairness assessment tool are open-source tools to make the adoption of responsible AI practices easier.
There’s also been new capabilities added in Vision Services and the Florence foundation model in Azure Cognitive Services. With these new capabilities, developers are able to improve content discoverability through automatic captions and video summaries. Features such as background removal, and retrieval of images are able to be fully automated. Users can also track movements and analyze environments in real-time.
To learn more check out the Azure Open Source Day Blog.