ax-platform 0.3.7

Adaptive Experimentation

Stars: 2270, Watchers: 2270, Forks: 292, Open Issues: 111

The facebook/Ax repo was created 5 years ago and the last code push was an hour ago.
The project is very popular with an impressive 2270 github stars!

How to Install ax-platform

You can install ax-platform using pip

pip install ax-platform

or add it to a project with poetry

poetry add ax-platform

Package Details

Author

Facebook, Inc.

License

MIT

Homepage

https://github.com/facebook/Ax

PyPi:

https://pypi.org/project/ax-platform/

GitHub Repo:

https://github.com/facebook/Ax

Classifiers

Related Packages

Errors

A list of common ax-platform errors.

Code Examples

Here are some ax-platform code examples and snippets.

GitHub Issues

The ax-platform package has 111 open issues on GitHub

• Scheduler + MapMetric test case that somewhat resembles IGML setup
• Created new context manager to mock expensive MBO functions
• BOTORCH_MODULAR with Surrogate(SingleTaskGP) doesn't support ax.plot.feature_importances.plot_feature_importance_by_feature, is ARD kernel still used?
• unsqueeze the tensor before evaluating acquisition function
• ModelBridge.evaluate_acquisition_function returns a length-one list regardless of how many observation features were specified
• Update input_constructors and re-enable KG/MES for modular BotAx
• Incorporate domain knowledge in categorial parameters
• Use model to calculate expected improvement for a list of candidates
• Logs contain warnings about transform_configs argument to TorchModelBridge containing an OptimizationConfig value, not expected by typecheck
• [FEATURE REQUEST] modify Ax API to allow for callable that evaluates a constraint and is passed to the optimizer
• Initializing Experiment with known data for Multi-Objective Optimization
• Generation strategy is not associated with correct experiment when using SQL storage after JSON deserialization
• Do you support discrete/mixed multi-objective optimization?
• Merge MultiObjectiveTorchModelBridge into TorchModelBridge.
• How to fix the random seed when using REMBO, ALEBO, HeSBO

See more issues on GitHub