SigOpt API Modules

SigOpt offers two API modules: Core Module and AI Module.

Core Module

The core module is designed for general usage of sample-efficient optimization. It has a lightweight API to give the users maximum flexibility. Users have the ability to create/delete/update almost all API objects throughout the optimization process.

The core module is optimization centric. It is most suitable for simulation optimization, configuration optimization, or general blackbox optimization problems. It is supported in multiple languages: Python, Java, and Bash. In particular, Bash users can access the API without installing a client library.

All core module specific documentation can be found under the CORE MODULE API REFERENCES section.

AI Module

The AI module is design for AI/ML use cases. In particular, one of its major features is the ability to track ML training runs in details, such as learning curve, metadata, model artifacts, etc. The AI module has integration with ML libraries and the ability to orchestrate your cluster.

The AI module is only available in Python.

All AI module specific documentation can be found under the AI MODULE API REFERENCES section.

Choose a Module

In the table below, we provide a more detailed list of comparison of the two modules to help you make an informed decision.

Features	Core Module	AI Module
Primary Use Cases	Simulation optimization, configuration optimization, general blackbox optimization	Machine learning (ML) hyperparameter tuning
Web Dashboard	Optimization focused	Project focused, with individual ML training run dashboard
APIs	Flexible API endpoints and objects design for general usage	API objects designed for AI/ML usage
Update/Delete all API objects	Supported at any point during an optimization experiment, via the API clients or the web dashboard	Partially supported for some objects, only through the web dashboard.
Advanced Experimentation Features	All supported	All supported
Supported Language	Python, Bash (Curl), Java, R, MATLAB	Python (including Jupyter notebook magic commands)
ML Training Run Tracking	N/A	Supported
ML Library Integration	N/A	XGBoost
Cluster Orchestration	N/A	SigOpt Orchestrate