Auto-PyTorch
automl.github.io
What it is
Auto-PyTorch is a free, self-hosted machine-learning toolkit based on PyTorch. It automates algorithm selection and hyperparameter tuning, including neural-network architecture and training settings, within a user-defined search budget. Documented search methods include Bayesian optimisation, meta-learning, SMAC and Hyperband. Supported work includes tabular classification, tabular regression and time-series forecasting. Tabular preprocessing can perform imputation, categorical encoding, scaling and feature processing, while related hyperparameters are tuned. The toolkit can build ensembles from validation-set predictions, with controls for ensemble size and candidate limits. Users can set estimator memory limits and total wall time. Parallel Bayesian optimisation uses Dask.distributed; workers need shared filesystem access to training data and models. Installation documentation specifies Linux, Python 3.7 or later, a C++11-capable compiler and SWIG 3.0. SWIG 4.0 or later is unsupported. Forecasting needs extra dependencies. A Docker image and PyPI installation are documented, under the 3-clause BSD license.
Spec sheet
- What it does
- Auto-PyTorch is an automated machine-learning toolkit based on PyTorch that helps users automate algorithm selection and hyperparameter tuning.Read 3 Oct 2026 · automl.github.io
- Purpose
- Auto-PyTorch is an automated machine learning toolkit based on PyTorch that automates algorithm selection and hyperparameter tuning.Read 3 Oct 2026 · automl.github.io
- Deployment
- The project provides a Docker image and can be installed from PyPI or manually in a Python environment.Read 3 Oct 2026 · automl.github.io
- Integrations
- The documentation describes using Dask.distributed for parallel Bayesian optimization and sklearn column transformers for data preprocessing.Read 3 Oct 2026 · automl.github.io
- License
- The documentation states that Auto-PyTorch is licensed under the 3-clause BSD license.Read 3 Oct 2026 · automl.github.io
- Maker
- The GitHub project says Auto-PyTorch is developed by the AutoML Groups of the University of Freiburg and Hannover.Read 3 Oct 2026 · github.com
- Compatibility limit
- The installation documentation says SWIG 4.0 or later is not supported.Read 3 Oct 2026 · automl.github.io
- Data preparation
- For tabular tasks, its preprocessing includes imputation, categorical encoding, scaling, and feature preprocessing, with corresponding hyperparameters tuned during search.Read 3 Oct 2026 · automl.github.io
- Ensembling
- It builds ensembles by selecting among models based on their predictions for a validation set, and users can configure ensemble size and candidate limits.Read 3 Oct 2026 · automl.github.io
- Forecasting dependencies
- Time-series forecasting requires additional dependencies beyond the base installation.Read 3 Oct 2026 · automl.github.io
- Installation
- The installation documentation specifies Linux, Python 3.7 or later, a C++11-capable compiler, and SWIG 3.0.*, and also documents a Docker image.Read 3 Oct 2026 · automl.github.io
- Integration
- The documented ecosystem includes PyTorch, scikit-learn transformers, Dask.distributed, and threadpoolctl.Read 3 Oct 2026 · automl.github.io
- Optimization
- It jointly optimizes neural network architecture and training hyperparameters for automated deep learning.Read 3 Oct 2026 · github.com
- Optimization methods
- It uses Bayesian optimization, meta-learning, and ensemble construction to search for models.Read 3 Oct 2026 · automl.github.io
- Parallel computing requirement
- When using multiple workers, the documentation says they must have access to a shared file system for training data and models.Read 3 Oct 2026 · automl.github.io
- Parallel processing
- It supports parallel Bayesian optimization using Dask.distributed, and parallel workers need access to a shared file system for training data and models.Read 3 Oct 2026 · automl.github.io
- Platform requirement
- The installation documentation lists Linux, Python 3.7 or later, a C++11-capable compiler, and SWIG 3.0 as system requirements.Read 3 Oct 2026 · automl.github.io
- Resource controls
- Users can set a memory limit for estimators and a total wall-time limit for model search.Read 3 Oct 2026 · automl.github.io
- Search methods
- Its search process uses Bayesian optimization, meta-learning, SMAC, and Hyperband to explore pipeline configurations within a user-set budget.Read 3 Oct 2026 · github.com
- Support and contribution
- The project invites bug reports, documentation improvements, and feature contributions through its GitHub issue tracker.Read 3 Oct 2026 · automl.github.io
- Support and contributions
- The project invites bug reports and documentation contributions through its GitHub issue tracker and recommends contacting developers by opening an issue before starting feature work.Read 3 Oct 2026 · automl.github.io
- Supported tasks
- The toolkit supports tabular classification, tabular regression, and time series forecasting.Read 3 Oct 2026 · github.com
Plans and prices
Auto-PyTorch Free 3-clause BSD license Read 3 Oct 2026 · automl.github.io
Strong points and limits
Strong points
- Free under a 3-clause BSD license
- Automates model selection and tuning
- Supports classification, regression and forecasting
- Configurable memory and wall-time limits
- Docker image is documented
Limits
- Linux installation is required
- SWIG 4.0 or later is unsupported
- Forecasting needs extra dependencies
- Parallel workers need shared storage
Who it suits
It suits developers and machine-learning practitioners working in Linux, self-hosted Python environments. Users planning parallel searches need shared filesystem access.
Verdict
Auto-PyTorch provides code-based automated modelling across several task types, with explicit resource controls and ensemble options. Its Linux prerequisites and shared-storage requirement shape deployment choices.
Where Specifiction places it
- AutoML Software#5 of 308.4
