From 977f45c25255cfded679a4c1b50617b181588024 Mon Sep 17 00:00:00 2001 From: Mauro Rigo Date: Fri, 17 Jan 2025 13:26:49 +0100 Subject: [PATCH] Updated README --- README.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/README.md b/README.md index c080f1b..06d41d7 100644 --- a/README.md +++ b/README.md @@ -1,12 +1,12 @@ -# jaxLDL: JAX implementation of Lagrangian Deep Learning +# JERALD: high-fidelity dark matter, stellar mass and neutral hydrogen maps from fast N-body simulations * [**Installation**](#installation) * [**Testing**](#testing) * [**Documentation**](#documentation) -jaxLDL is the [JAX](https://github.com/google/jax) implementation of the [code](https://github.com/biweidai/LDL) performing [Lagrangian Deep Learning](https://arxiv.org/abs/2010.02926), a machine learning method to paint baryons and related quantities on top of N-body only simulations. +JERALD is a code, based on the [Lagrangian Deep Learning](https://arxiv.org/abs/2010.02926) idea of Dai and Seljak, to paint baryons and related quantities on top of N-body only simulations with machine learning. -The package is based on JAX and [mpi4jax](https://github.com/mpi4jax/mpi4jax/tree/master), and it uses MPI to implement parallel Cloud-In-Cell (CIC) painting and interpolation algorithms as well as parallel forward and backward Fourier transforms via [FFTW](https://www.fftw.org/). +The package is based on [JAX](https://github.com/google/jax) and [mpi4jax](https://github.com/mpi4jax/mpi4jax/tree/master), and it uses MPI to implement parallel Cloud-In-Cell (CIC) painting and interpolation algorithms as well as parallel forward and backward Fourier transforms via [FFTW](https://www.fftw.org/). ## Installation To install the package, clone the repo and run: @@ -20,7 +20,7 @@ CUDA support isn't required, as for the time being the package only runs on CPU. On a cluster, I suggest loading any MPI modules and then installing mpi4jax via pip, to allow mpi4py to build on the existing MPI installation. ## Testing -A simple test that computes the LDL loss and its derivative with respect to the LDL parameters (checked against the original code) is available in ```losstest.py```. To run it with a single MPI process, execute +A simple test that computes the model loss and its derivative with respect to its parameters (checked against the original LDL code) is available in ```losstest.py```. To run it with a single MPI process, execute ``` python losstest.py ```