Update README.md

README.md

@@ -2,6 +2,10 @@
 
 Convolutional neural network used to deconvolute beamforming maps.
 
+This work is a part of the POLA3 project.
+
+Wagner Gonçalves Pinto
+
 <table border="1" align="center">
   <thead>
     <tr>
@@ -45,12 +49,10 @@ Convolutional neural network used to deconvolute beamforming maps.
 </table>
 
 
-This work is a part of the POLA3 project.
-
-Wagner Gonçalves Pinto
-
 ## Installing dependencies
 
+Neural network is implemented using frameworks [PyTorch](https://pytorch.org/) (version 1.8.0) and [PyTorch Lightning](https://www.pytorchlightning.ai/) (1.2.4).
+
 File `conda_env.yml` contains the list and versions of the packages used by the library.
 
 To install all dependencies, one may use [conda](https://anaconda.org/):
@@ -61,9 +63,9 @@ conda env create -f conda_env.yml --verbose
 
 ## Usage
 
-Scripts `generate.py`, `train.py` and `test.py` provide interfaces to performing database generation, neural network training and testing from a shell script. Example of calls and a description of the flags are presented next.
+Folder [deconvnetlib](deconvnetlib) contains the DeconvNet implementation and scripts associated with the database generation and loading. Neural network architecture is defined in [deconvnetlib/network.py](deconvnetlib/network.py). Folder [beamlib](beamlib) is a git-submodule of a library used for beamforming.
 
-Neural network architecture is defined in [deconvnetlib/network.py](deconvnetlib/network.py)
+Scripts `generate.py`, `train.py` and `test.py` provide interfaces to performing database generation, neural network training and testing from a shell script. Example of calls and a description of the flags are presented next. More elaborate scripts are available in the [templates](templates) folder.
 
 ### Generating the database