tsinghua-fib-lab/UniST
Official implementation for "UniST: A Prompt-Empowered Universal Model for Urban Spatio-Temporal Prediction" (KDD 2024)
This project helps urban planners, transportation managers, and city officials predict various urban spatio-temporal events like traffic flow, crowd density, or taxi demand. It takes historical spatio-temporal data from multiple cities and domains as input and outputs predictions for short-term, long-term, few-shot, and even entirely new scenarios. This is ideal for those who need to forecast urban dynamics accurately and efficiently.
213 stars. No commits in the last 6 months.
Use this if you need a flexible model to predict urban phenomena across different cities and scenarios, especially when data is limited or when generalizing to unseen situations.
Not ideal if your prediction tasks are not related to urban spatio-temporal data or if you prefer to train separate, highly specialized models for each specific prediction problem.
Stars
213
Forks
22
Language
Python
License
—
Category
Last pushed
Jan 09, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/tsinghua-fib-lab/UniST"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
amazon-science/chronos-forecasting
Chronos: Pretrained Models for Time Series Forecasting
SalesforceAIResearch/uni2ts
Unified Training of Universal Time Series Forecasting Transformers
moment-timeseries-foundation-model/moment
MOMENT: A Family of Open Time-series Foundation Models, ICML'24
ServiceNow/TACTiS
TACTiS-2: Better, Faster, Simpler Attentional Copulas for Multivariate Time Series, from...
yotambraun/APDTFlow
APDTFlow is a modern and extensible forecasting framework for time series data that leverages...