PyMC
Bayesian modeling with PyMC: priors, NUTS sampling and model comparison.
Install
Ships scriptsgh skill install K-Dense-AI/scientific-agent-skills pymc
Building a Bayesian or hierarchical model, running MCMC or variational inference, or debugging divergences and convergence.
- Author
- K-Dense Inc.
- License
- Apache-2.0
Replaces the habit of sampling first and checking afterwards with a fixed order: a prior predictive check before fitting, then R-hat, effective sample size and divergence counts before anything is read off the posterior. Divergences are treated as a modeling problem to reparameterize, not noise to bury by raising target_accept. Hierarchical structure, missing data and measurement error are presented as modeling decisions, and candidates are compared with LOO and WAIC. Two scripts handle diagnostics and comparison; examples target PyMC 6.0.1 on PyTensor 3.
Similar skills
Performance Optimization
SkillProfile first, fix the measured bottleneck, then measure again.
pyOpenMS
SkillMass spectrometry pipelines for proteomics and metabolomics via OpenMS.
RAG Implementation
SkillBuilds retrieval pipelines that ground answers in real documents.
Comments(0)
Sign in to comment