# IBM Granite releases an ensemble of four models for time-series forecasting

- Published: 2026-10-22T07:51:31.000Z
- Updated: 2026-10-22T10:03:02.732Z
- Section: Models
- Page: https://aidailyjournal.com/en/n/2026/10/22/ibm-granite-releases-an-ensemble-of-four-models-for-time-series-forecasting/
- Sources: [Hugging Face](https://huggingface.co/blog/ibm-research/granite-timeseries-ensemble-r1)

## Summary

IBM Granite presents Granite Timeseries Ensemble, an extensible interface that combines four pretrained models with different architectures: Granite PatchTST-FM-r1, Granite PatchTST-FM-r2, Granite FlowState-r1.1, and Granite TTM-r3. The recommended recipe, called Probabilistic Ensemble, gives each model equal weight and combines their forecasts into a single result without retraining or fitting weights for each series; an alternative uses the interquartile range (IQR) to combine the models adaptively.

The tool returns a point forecast and quantile forecasts, which by default range from 0.1 to 0.9. The 0.1 and 0.9 quantiles can represent a central interval with nominal 80% coverage. The configuration is available on Hugging Face under a permissive license and downloads the four models’ checkpoints during the first run.

## Why it matters (editorial interpretation)

The interface provides a ready-made way to combine point forecasts and uncertainty intervals from four models without fitting weights for each series. This could simplify experimentation with probabilistic forecasting methods, although the 80% coverage is nominal rather than a performance guarantee for every case.

_Summary written by AI Daily Journal in its own words, based on the sources above. For details, read the original articles._