TimeNet: Open-source data infrastructure for Temporal Foundation Models

Hi HN,<p>Current frontier LLMs show incredible results across text, video, and audio, but don&#x27;t do as well with real-world scenarios like production lines, trading, drilling operations, or demand supply forecasting. They are not capable of comprehending this data, especially given a limited context.<p>Models exist that are able to understand temporal information natively—Temporal Foundation Models (TFMs). However, there is a data problem: unlike all the public information that frontier models train on, temporal information is still scattered across organizations without a unified way to access it.<p>That’s why we built TimeNet. It&#x27;s like ImageNet for temporal data:<p>- One uniform data layer for either our repository or your own data<p>- Define tasks on top temporal data around classification, question answering, scalar prediction, temporal localization, forecasting, generation, editing, and correspondence (more to come)<p>- Simple and performant API to integrate into ML frameworks.<p>- First experimental results: training on multiple temporal datasets increased F1 performance on each dataset respectively by an average of 14% and 22% for OpenTSLM and SLIP respectively.<p>- Install via ‘uv add timenet’<p>TimeNet is the result of a growing open and global collaboration between researchers from Aionic Labs, Stanford, ETH Zurich, UNSW Sydney, Sun Yat-sen University, Griffith University, Google, UCLA, TimeCopilot, Dartmouth, Harvard, Cambridge, Eindhoven University of Technology and KAIST.<p>Paper: <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2610.04407" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2610.04407</a><p>Quickstart: <a href="https:&#x2F;&#x2F;docs.timenet.ai&#x2F;get-started.html" rel="nofollow">https:&#x2F;&#x2F;docs.timenet.ai&#x2F;get-started.html</a><p>TimeNet Registry: <a href="https:&#x2F;&#x2F;registry.timenet.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;registry.timenet.ai&#x2F;</a><p>There is currently an online hackathon ongoing on TimeNet (<a href="https:&#x2F;&#x2F;challenge.timenet.ai&#x2F;" rel="nofollow">https:&#x2F;&#x2F;challenge.timenet.ai&#x2F;</a>) with up to $50k in cloud credit prizes, this should fuel your first model training. We would love to see you build on top of it and to hear your thoughts!