Update README.md

README.md

@@ -15,20 +15,19 @@ This is the official repository for the paper [Iterated Denoising Energy Matchin
 
 We propose iDEM, a scalable and efficient method to sample from unnormalized probability distributions. iDEM makes use of the DEM objective, inspired by the stochastic regression and simulation
 free principles of score and flow matching objectives while allowing one to learn off-policy, in a loop while itself generating (optionally exploratory) new states which are subsequently
-learned on. iDEM is also capable of incorporating symmetries, namely those represented by the product group of $SE(3) \\times \\mathbb{S}\_n$. We experiment on a 2D GMM task as well as a number of physics
-inspired problems. These include:
+learned on. iDEM is also capable of incorporating symmetries, namely those represented by the product group of $SE(3) \\times \\mathbb{S}\_n$. We experiment on a 2D GMM task as well as a number of physics-inspired problems. These include:
 
-- DW4 -- the 4 particle double well potential (8 dimensions total)
-- LJ13 -- the 13 particle Lennard-Jones potential (39 dimensions total)
-- LJ55 -- the 55 particle Lennard-Jones potential (165 dimensions total)
+- DW4 -- the 4-particle double well potential (8 dimensions total)
+- LJ13 -- the 13-particle Lennard-Jones potential (39 dimensions total)
+- LJ55 -- the 55-particle Lennard-Jones potential (165 dimensions total)
 
-This code was taken from an internal repository and as such all commit history is lost here. As such, development credit for this repository goes primarily to [@atong01](https://github.com/atong01), [@jarridrb](https://github.com/jarridrb) and [@taraak](https://github.com/taraak) who built
+This code was taken from an internal repository and as such all commit history is lost here. Development credit for this repository goes primarily to [@atong01](https://github.com/atong01), [@jarridrb](https://github.com/jarridrb) and [@taraak](https://github.com/taraak) who built
 out most of the code and experiments with help from [@sarthmit](https://github.com/sarthmit) and [@msendera](https://github.com/msendera). Finally, the code is based off the
 [hydra lightning template](https://github.com/ashleve/lightning-hydra-template) by @amorehead and makes use of the [FAB torch](https://github.com/lollcat/fab-torch) code for the GMM task and replay buffers.
 
 ## Installation
 
-For installation we recommend the use of Micromamba. Please refer [here](https://mamba.readthedocs.io/en/latest/installation/micromamba-installation.html) for an installation guide for Micromamba.
+For installation, we recommend the use of Micromamba. Please refer [here](https://mamba.readthedocs.io/en/latest/installation/micromamba-installation.html) for an installation guide for Micromamba.
 First, we install dependencies
 
 ```bash
@@ -55,7 +54,7 @@ To run an experiment, e.g., GMM with iDEM, you can run on the command line
 python dem/train.py experiment=gmm_idem
 ```
 
-We include configs for all experiments matching the settings we used in our paper for both iDEM and pDEM with the exception of LJ55 for
+We include configs for all experiments matching the settings we used in our paper for both iDEM and pDEM except LJ55 for
 which we only include a config for iDEM and pDEM had convergence issues on LJ55.
 
 ## Current Code