bayesflow-simulator

What problem does it solve?

This Skill assists in the creation, modification, and debugging of data-generating processes (simulators) for the BayesFlow 2.x library, ensuring correct implementation of prior, likelihood, and meta functions.

Core Features & Use Cases

• Simulator Factory: Provides patterns for both subclassing Simulator (for variable-size data) and using the make_simulator() function (for fixed-size data).
• Convention Enforcement: Guides users on critical aspects like function signatures, RNG discipline, output dictionary conventions, and configuration practices.
• Use Case: You need to build a new simulator for a complex statistical model. This Skill will guide you through defining the prior distributions, the likelihood function, and any necessary meta-parameters, ensuring the simulator integrates seamlessly with BayesFlow.

Quick Start

Use the bayesflow-simulator skill to create a new simulator using the make_simulator function with provided prior, likelihood, and meta functions.