Azure AI foundry provides an accessible and hands on way to deploy your own custom large language model AI applications based of models such as chat GPT. AI foundry allows you to customise how the model functions, as well as giving you the opportunity to train models on and integrate them with your own documents and data. These specific models can then be accessed through AI foundry’s API, allowing you to embed them into your own applications.
Unlike the tools such ML studio and the computer vision tools in AI services, AI foundry allows you to design a models or agents with the same knowledge base that can be accessed by multiple users across your research group or organisation. This can make it easy to create research specific models for your research groups which are based on your chosen collection of academic literature and similar sources, creating more robust and intelligent research tool.
AI foundry has a wide range of the most up to date large language models available, and with each model assessed using intuitive metrics, it makes it easy to pick the right models for you. AI foundry also allows you to customise the throughput of the model, allowing you to easily scale the cost of the model with your needs.