From 77cf246d49b05d9314a137f30da6d8cdc0ebef3a Mon Sep 17 00:00:00 2001 From: dom0015 Date: Tue, 28 May 2024 21:23:46 +0000 Subject: [PATCH] Readme change --- README.md | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 96e2575..8b421c9 100644 --- a/README.md +++ b/README.md @@ -32,7 +32,10 @@ $f_{U|Y}\left(u|y\right)\propto{f_{Z}\left(y-G\left(u\right)\right)}{f_{U}\left( - a solver is typically a linked numerical library - number of solvers is typically lower than number of chains -## Usage: +## Getting started: + - open repository in the Docker container (e.g. using the Dev Containers extension of Visual Studio Code) + - `pip install .` + Before running the sampling process, it is necessary to specify: - configuration (basic settings, e.g. number of solvers, initial samples, ...) - prior @@ -41,7 +44,9 @@ Before running the sampling process, it is necessary to specify: - surrogate model updater - list of stages -See examples in the **/toy_examples** folder. +See examples in the **toy_examples** folder, e.g.: + - `cd toy_examples/` + - `mpiexec -n 4 python3 -m mpi4py minimal_example.py` ## Surrogate models: The following non-intrusive surrogate models are implemented. They are constructed from snapshots $\left(u^{\left(k\right)},G\left(u^{\left(k\right)}\right)\right)$ and they can be adaptively refined during the sampling process.