AutoGluon
auto.gluon.ai
- Free plan
- Windows
- macOS
- Linux
- Self-hosted
What AutoGluon is
AutoGluon is a free, open-source Python library for automating machine learning on tabular data, time series, and multimodal tasks. For tables, it trains classification and regression models while taking care of data cleaning, feature engineering, hyperparameter optimization, and model selection. Its TimeSeriesPredictor creates probabilistic forecasts for multiple series from historical data and related covariates. AutoMM reduces manual work when adapting foundation models to specialized data. Documentation examples cover image and text prediction, object detection, image segmentation, document prediction, entity extraction, and semantic matching. AutoMM can adapt models from Hugging Face, TIMM, and MMDetection collections; forecasting methods include ETS, ARIMA, LightGBM, DeepAR, Temporal Fusion Transformer, and Chronos. AutoGluon requires Python 3.10 through 3.13 and supports Linux, macOS, and Windows. GPU use is not supported on macOS. For cloud training and deployment, the project points to AutoGluon Cloud and deep learning containers; SageMaker compute and S3 storage are billed to the user's AWS account. The library is licensed under Apache 2.0.
Who it suits
AutoGluon suits Python users building predictive models from tables, time series, images, text, or other supported data. It may fit teams that want automated model selection or forecasting and are prepared to work in a supported Python environment.
Good to know, and what to watch
- Automates data cleaning and feature engineering for tabular models.
- Builds probabilistic forecasts across multiple time series.
- Documentation covers image, text, and document prediction tasks.
- Free and licensed under Apache 2.0.
- GPU use is not supported on macOS.
- Requires Python 3.10 through 3.13.
- SageMaker compute and S3 storage are billed separately.
Plans
Every plan and priceAutoGluon
Open source
Getting started
- Install Python 3.10 through 3.13
- Install AutoGluon on Linux, macOS, or Windows
- Use the library for tabular, time-series, or multimodal modeling
- Use Linux or Windows when GPU processing is required
- Run it self-hosted or use AutoGluon Cloud and deep learning containers
- Use the pre-installed package in Amazon SageMaker Distribution
Limits to check
The library is free, but GPU processing is not supported on macOS. When using SageMaker, compute and S3 storage are billed to the user’s AWS account.
The verdict
AutoGluon covers tabular prediction, time-series forecasting, and a broad set of multimodal examples in one Python library. Check its Python requirements and macOS GPU limitation, and account for AWS compute and storage charges when using SageMaker.
Questions people ask
Does AutoGluon cost anything?
The AutoGluon library is free and open source under the Apache 2.0 license. AutoGluon Cloud is described as a free wrapper, while SageMaker compute and S3 storage are billed separately.
Which operating systems and Python versions are supported?
It supports Linux, macOS, Windows, and self-hosted use, with Python versions 3.10 through 3.13.
Can AutoGluon use a Mac GPU?
No. GPU usage is not supported on macOS; the installation guidance directs GPU users to Linux or Windows.
What machine-learning tasks does it cover?
It supports tabular classification and regression, time-series forecasting, image and text prediction, object detection, segmentation, document prediction, entity extraction, and semantic matching.
Is AutoGluon available in SageMaker?
AutoGluon is pre-installed in all Amazon SageMaker Distribution releases. SageMaker Autopilot is described as a managed AutoGluon experience.
What the maker publishes
- Cloud cost
- AutoGluon-Cloud is described as a free wrapper, while SageMaker compute and S3 storage are billed to the user's AWS account. source
- Deployment
- The project points to AutoGluon Cloud and deep learning containers for cloud training and deployment, and describes SageMaker Autopilot as a managed AutoGluon experience. source
- Forecasting models
- Time-series forecasting combines methods including ETS and ARIMA from StatsForecast, LightGBM, DeepAR and Temporal Fusion Transformer from GluonTS, and Chronos. source
- Founded
- 2019 source
- GPU limitation
- GPU usage is not supported on macOS; the installation guide directs users to Linux or Windows for GPU use. source
- Install requirements
- AutoGluon requires Python 3.10 through 3.13 and supports Linux, macOS, and Windows. source
- Integrations
- AutoMM adapts models from Hugging Face, TIMM, and MMDetection model zoos. source
- License
- The AutoGluon library is licensed under Apache 2.0. source
- Multimodal
- AutoMM reduces manual work in preprocessing, model selection, and fine-tuning to adapt foundation models to domain-specific data. source
- Purpose
- AutoGluon automates machine learning for data such as tables and time series, helping users build predictive models with a few lines of code. source
- SageMaker
- AutoGluon is pre-installed in all releases of Amazon SageMaker Distribution. source
- Security reporting
- The contribution guide asks users to report potential security issues to AWS/Amazon Security through its vulnerability reporting page rather than filing a public GitHub issue. source
- Support
- The project welcomes bug reports and feature requests through its GitHub issue tracker. source
- Supported tasks
- The documentation includes image and text prediction, object detection, image segmentation, document prediction, entity extraction, and semantic matching examples. source
- Tabular modeling
- AutoGluon can train classification and regression models from tabular data while handling data cleaning, feature engineering, hyperparameter optimization, and model selection. source
- Time series
- Its TimeSeriesPredictor trains multiple models to generate probabilistic forecasts for multiple time series using historical data and related covariates. source
Checked against the maker's pages on 3 Oct 2026.