What is the correct order for BayesFlow adapter operations?

The pipeline should follow this sequence: Type conversion, then cross-key transforms, followed by per-key transforms (like log or expand_dims), and finally standardization before concatenating into schema slots.

What is a BayesFlow adapter?

A BayesFlow adapter is a specialized data preprocessing pipeline that transforms raw simulator outputs into the specific dictionary schema (inference, summary, and conditions) required by neural inference networks.

Why should I use prior moments instead of empirical statistics for standardization?

Empirical moments can shift across training batches, especially with variable-size data, causing instability. Analytical prior moments are deterministic and constant, preventing train/validation distribution shifts.

How do I handle shape mismatches during concatenation?

You should use the `expand_dims` method within the adapter to add necessary axes (e.g., axis=-1) to parameters before attempting to concatenate them into schema variables.

BayesFlow Adapter Assistant

Name: BayesFlow Adapter Assistant
Author: matthiaskloft

bymatthiaskloft

0•

数据科学与机器学习

Streamlines the creation and debugging of BayesFlow data preprocessing pipelines for simulation-based inference.

This skill provides Claude with domain-specific expertise in BayesFlow adapters, ensuring data preprocessing pipelines correctly map simulator outputs to the required inference, summary, and condition schemas. It enforces critical best practices such as utilizing analytical prior moments for standardization to prevent distribution shifts and ensuring the correct pipeline order where transforms are applied before standardization. Whether you are implementing complex reparameterizations, handling broadcasting for set-level parameters, or debugging shape mismatches, this skill provides the patterns and validation logic needed for robust amortized Bayesian inference.

主要功能

01Analytical prior moment standardization to prevent training distribution shifts

02Declarative AdapterSpec patterns for HPO-compatible workflows

03Schema-aware mapping for inference, summary, and condition slots

04Enforcement of transform-before-standardize pipeline ordering

050 GitHub stars

06Support for custom elementwise and cross-key reparameterization transforms

使用场景

01Building data pipelines for amortized Bayesian inference models

02Implementing complex parameter reparameterizations within BayesFlow

03Debugging shape mismatches and broadcasting errors in simulator outputs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add matthiaskloft/claude-skills bayesflow-adapter

For use in Claude.ai and ChatGPT

主要功能

01Analytical prior moment standardization to prevent training distribution shifts

02Declarative AdapterSpec patterns for HPO-compatible workflows

03Schema-aware mapping for inference, summary, and condition slots

04Enforcement of transform-before-standardize pipeline ordering

050 GitHub stars

06Support for custom elementwise and cross-key reparameterization transforms

使用场景

01Building data pipelines for amortized Bayesian inference models

02Implementing complex parameter reparameterizations within BayesFlow

03Debugging shape mismatches and broadcasting errors in simulator outputs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add matthiaskloft/claude-skills bayesflow-adapter

For use in Claude.ai and ChatGPT