Understand the project
Practical overview
When it fits
Use Chronos when you need pretrained time-series forecasting models to incorporate into a Python workflow, and are prepared to build or connect the surrounding application or deployment process.
Who it is for
- Python developers adding pretrained forecasting to a time-series workflow.
- Teams comparing model variants for univariate, multivariate or covariate-informed forecasting.
Documented uses
- Generate forecasts from historical time-series observations.
- Explore zero-shot forecasting with Chronos-2, including tasks with multiple series or covariates.
- Choose among model checkpoints with different parameter sizes and forecasting approaches.
Before you choose
Requirements & limits
- Chronos-2 supports zero-shot univariate, multivariate and covariate-informed forecasting.
- Original Chronos models generate probabilistic forecasts by sampling future trajectories; Chronos-Bolt generates quantile forecasts directly across multiple future steps.
- Available checkpoints range from 8 million to 710 million parameters.
- The repository identifies Python package installation as the inference entry point and links to Hugging Face model collections and a SageMaker deployment tutorial.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Source-based practical guide
Install the Chronos forecasting package
Set up the Python package before following a model-specific inference example.
Before you start
- Python environment with pip
Steps
Install the package with the command below.
pip install chronos-forecastingContinue with the repository's Chronos-2 quickstart notebook for a model-specific starting point.
Expected outcome
The Chronos package is installed, and the repository identifies a quickstart notebook for beginning Chronos-2 inference.
Guide limitations
- The documented installation step alone does not run a forecast.
- The quickstart notebook's execution requirements and deployment details are not included here.
Evidence and transparency
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 7 October 2026
- Original website title
- amazon-science/chronos-forecasting · Official homepage title · retained in its source language