Yug: a foundation model for zero-shot probabilistic time-series forecasting
-
Updated
Oct 9, 2026 - Python
Yug: a foundation model for zero-shot probabilistic time-series forecasting
First Apache-2.0 Turkish-vertical TSFM fine-tune: Chronos-2 + LoRA r=32 + V4 router (MASE-aware adoption) + bucket isotonic calibration. GIFT-Eval submission 2026-05-27.
Fracast-0 is an 85K-parameter pretrained Time Series Foundation Model, small enough to load directly in a web page.
A systematic robustness testing tool for time series foundation models in edge cases.
Foundation model evaluation on pupillary light reflex time series. Biosignal analysis with modern TSFMs.
A grounded time-series forecasting framework that combines a language controller with specialist forecasting models. It turns natural-language requests into structured forecasting plans and evidence-grounded predictions, substantially outperforming general-purpose conversational language models used as direct forecasting baselines.
Paper's code of "How Effective Are Time Series Foundation Models for Epidemiological Data Analysis?"
InfraWatch: Time series foundation model anomaly detection for infrastructure metrics. Zero-config, maintenance-window aware.
MCP server for time-series foundation model (TSFM) forecasting: probabilistic forecasts from Chronos-2, TimesFM 2.5, Toto 2, TiRex-2 and IBM Granite models, or an accuracy-weighted ensemble.
Research and operations platform for statistically rigorous lottery forecasting experiments across six games.
To associate your repository with the tsfm topic, visit your repo's landing page and select "manage topics."