🚧 This repository is under construction. 🚧 Expect a public release coming this Summer 2022.

Want to contribute? Check out the GitHub container repository Limboid/the-artificial-ecosystem for this project.

Modality primitives for building modality agnostic models and training pipelines

• Provide minimally descriptive unambiguous representation of information structure
• Support structures: flat, set, sequence, grid, and graph
• Support primitive types: boolean, categorical, range, integer, string, and real
• Support arbitrary tags: context, human_label, learned_embedding, etc.
• make pypi and npm packages