IBM launches Granite Timeseries Ensemble multi-model integration interface
Current overview (not historical)
IBM has launched the Granite Timeseries Ensemble interface, providing a pre-trained model integration solution comprising four architectures: PatchTST-FM, Conformer, FlowState, and TTM. Users can generate point predictions and quantile predictions without training their own models. The default configuration averages the four models equally, while an IQR-based scheme dynamically adjusts weights based on prediction distribution width. Models are stored in configuration form and downloaded automatically on first run.
Written with AI from the reporting · updated 10/22/2026
Expand the detailed fact route and sources
Story route
A station-by-station route from the first public evidence onward; every station carries its own evidence and uncertainty.
- 1Unclassified fact stationOct 22, 00:00recorded Oct 22, 16:01single sourcerevised
发布 Granite Timeseries Ensemble 多模型集成预测接口
Read original source ↗ Hugging Face BlogReplay this moment in the explorerEvidence & revisions · 1 source records
- Subject：
- hugging-face
- Action：
- 发布 ensemble 接口
- Object：
- 时间序列预测模型集成
- Occurred (fact side)：
- Oct 22, 00:00
- Published (valid side)：
- Oct 22, 15:51
- First seen (known side)：
- Oct 22, 16:01
Fact stationSingle source · one-sided claimRevisedCopy may have been revised later; time replay is bounded. Layer labels group reading only, not causation between layers.
Uncertainty marks flag where to check the sources; they are not proven causation.
Report timeline
Explore every side through dated source reports.
- Hugging Face BlogFirst-party sourceRelated fact: single sourceSource report revisedIBM Granite introduces ensemble forecasting using four pre-trained time-series models
Report first recorded Oct 22, 16:01
IBM Granite Timeseries Ensemble combines four complementary pre-trained Granite architectures—PatchTST-FM-r1, PatchTST-FM-r2, FlowState-r1.1, and TTM-r3—to provide median and quantile forecasts without requiring dataset training or fitted ensemble weights. Users can start with the Probabilistic Ensemble recipe and extend the interface, offering point forecasts and quantile predictions (defaulting from 0.1 to 0.9) for applications requiring forecast ranges.
Read original source ↗
Event discussion
Reader opinions are hosted on GitHub Discussions; posting requires a GitHub account. Comments are not verified facts.