Can this skill help with slow machine learning tests?

Yes, it provides patterns for using MagicMock to simulate inference and summary networks, allowing you to test pipeline logic and shapes in seconds rather than minutes.

How does it handle non-deterministic test failures?

The skill enforces the use of seeded NumPy Generator (RNG) instances and discourages global state seeds, ensuring that every test run is reproducible and deterministic.

Why do I need to set the Keras backend in conftest.py?

Keras 3 reads the KERAS_BACKEND environment variable only once at the first import. This skill ensures the variable is set to 'torch' before any BayesFlow or Keras modules are loaded to prevent backend conflicts.

Does it support symmetry testing for neural networks?

Absolutely. It includes specific patterns for Invariance and Equivariance testing (IRT), ensuring that permuting input data produces the expected symmetric results.

BayesFlow Testing Suite

Name: BayesFlow Testing Suite
Author: matthiaskloft

bymatthiaskloft

0•

Seguridad y Pruebas

Standardizes and accelerates the creation of robust unit and integration tests for BayesFlow simulation-based inference extensions.

The BayesFlow Testing skill provides specialized guidance for building high-quality test suites for simulation-based inference tools. It addresses critical technical requirements such as the mandatory Keras 3 backend initialization in conftest.py, ensuring that 'torch' is configured before any BayesFlow imports. The skill includes advanced patterns for shape verification, configuration serialization checks, and mock-based testing to avoid slow training cycles. It is particularly useful for developers building complex neural architectures that require invariance or equivariance testing, helping to maintain scientific rigor and software stability in research-oriented codebases.

Características Principales

01Permutation invariance and equivariance verification patterns

02Shape-testing templates for simulators, networks, and adapters

03Critical Keras 3 backend setup automation for Torch compatibility

04Serialization and roundtrip validation for config dataclasses

05Mock-based inference network testing to speed up CI/CD

060 GitHub stars

Casos de Uso

01Debugging Keras backend initialization issues in BayesFlow extension packages

02Writing unit tests for custom simulators to ensure output shape consistency

03Implementing fast CI pipelines that verify network logic without full model training

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add matthiaskloft/claude-skills bayesflow-testing

For use in Claude.ai and ChatGPT

Características Principales

01Permutation invariance and equivariance verification patterns

02Shape-testing templates for simulators, networks, and adapters

03Critical Keras 3 backend setup automation for Torch compatibility

04Serialization and roundtrip validation for config dataclasses

05Mock-based inference network testing to speed up CI/CD

060 GitHub stars

Casos de Uso

01Debugging Keras backend initialization issues in BayesFlow extension packages

02Writing unit tests for custom simulators to ensure output shape consistency

03Implementing fast CI pipelines that verify network logic without full model training