AI4Compose

In this collection you will find the developments for the AI4Compose software solution (https://github.com/ai4os/ai4-compose/).

AI4Compose is a suite of software utilities designed to facilitate the integration and orchestration of multiple artificial intelligence (AI) models and machine learning techniques. Its primary function is to amalgamate various AI components into organised workflows or pipelines. AI4Compose integrates with well-known visual programming tools such as Elyra and Node-RED (along with FlowFuse), enabling the creation of workflows from the inference of pre-trained models. Users can work with nodes and connections between them. Despite their different functionalities and capabilities, both platforms share a common purpose in utilising graphical composition tools applicable to AI models. As a tool for executing inference tasks, AI4Compose has been integrated with OSCAR (https://oscar.grycap.net/), a serverless framework used to execute the inference phase of the AI models coming from the AI4EOSC Marketplace (https://dashboard.cloud.ai4eosc.eu/catalog/modules).

Specifically for Node-RED, AI4Compose introduces new modules based on subflows. These subflows efficiently employ standard Node-RED nodes to make requests to OSCAR services. Users can effortlessly interact with these modules in a graphical environment through a "drag and drop" interface, enabling them to compose complex workflows with ease. For more details, please visit the AI4Compose official documentation, available at: https://docs.ai4os.eu/en/latest/howtos/pipelines/flowfuse.html

The AI4EOSC project has received funding from the European Union’s Horizon Research and Innovation programme under Grant agreement No. 101058593