TimeNet: Open-source data infrastructure for temporal foundation models
Summary
TimeNet is an open-source data infrastructure designed to store and manage time-series datasets in a single TimeF format, enabling a unified API for finding, building, downloading, and loading data for temporal foundation models. The project provides a registry and connectors to bridge raw datasets to TimeF, while keeping model training and evaluation in the user’s chosen framework. TimeNet emphasizes standardized data access and interoperability for AI workloads.