#26 · Primary category: Time Series Machine Learning

uni2ts

deep-learning forecasting machine-learning pre-trained-models pre-training representation-learning time-series time-series-forecasting transformers universal-forecasting

Unified Training of Universal Time Series Forecasting Transformers

Project last updated:06/02/26

GitHub Stars

1.6K

Forks

214

Contributors

11

License

Apache-2.0

Why we included this project

Uni2TS is the PyTorch implementation behind Salesforce's Moirai line of pre-trained time series transformers, so it is the place to start if you want forecasts without training a model yourself. Its zero-shot checkpoints load from Hugging Face and produce predictions straight from a pandas DataFrame, which suits demand planning, inventory, or energy forecasting. The repo also covers large-scale pre-training and fine-tuning, and its evaluation code compares against models like Chronos, TimesFM, and VisionTS on standard benchmarks. Notebooks and worked examples make it useful both as deployable tooling and as a reference for how universal forecasting models are built and validated.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category