Run affine-invariant sampler in automated fashion until convergence.
Project description
autoemcee
Runs MCMC automatically to convergence.
About
Runs a family of Markov Chain Monte Carlo ensemble samplers (Affine-Invariant or Slice Sampler) with gradually increasing number of samples until they converge.
Convergence is tested for each ensemble and for each parameter with the Geweke diagnostic. Additionally, across ensembles, the Gelman-Rubin r-hat is used.
Can be run with MPI without modifying the code.
This package is built on top of emcee, zeus, anviz and mpi4py.
You can help by testing autoemcee and reporting issues. Code contributions are welcome. See the Contributing page.
Features
Pythonic. pip installable.
Easy to program for: Sanity checks with meaningful errors
Supports both emcee and zeus
MPI support
Usage
Read the full documentation at:
https://johannesbuchner.github.io/autoemcee/
For parallelisation, use:
mpiexec -np 4 python3 yourscript.py
Licence
GPLv3 (see LICENCE file). If you require another license, please contact me.
Release Notes
0.1.0 (2020-03-07)
First version
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for autoemcee-0.2.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7b78f56b222631be510f0d3bf1b08dc96d0cb96ad35bfde03daa2a0e83de3c22 |
|
MD5 | 015a4b6c27f44ab0975ac4c205e16da9 |
|
BLAKE2b-256 | 23be76bb732ec1887bf19a898771c083aae0d911598a22f24a37e6ed4fb4b73c |