diff --git a/README.md b/README.md index 80bc646b579daeae9136d29315785cb8c854b482..0c0a34f046bd4b05f21da00ed74a5bcc7aa6faa1 100644 --- a/README.md +++ b/README.md @@ -10,11 +10,7 @@ The above paper contains theoretical and numerical results that can be reproduce `srenet` (**S**creen & **R**elax method for the **E**lastic-**NET** problem) package works with Python 3.9+. Please refer to the [Python download page](https://www.python.org/downloads/) for install instructions. -Dependencies : -* [NumPy](http://www.numpy.org) -* [SciPy](https://www.scipy.org) -* [Matplotlib](http://matplotlib.org) -* [Pandas](https://pandas.pydata.org) +**Dependencies** : [NumPy](http://www.numpy.org), [SciPy](https://www.scipy.org), [Matplotlib](http://matplotlib.org), [Pandas](https://pandas.pydata.org) ## Install instructions @@ -47,8 +43,8 @@ pip install . ## ICASSP experiments Experiments presented in the Screen & Relax paper are located in the folder `screen-and-relax/exp/ICASSP/`. The two available experiments are -* `convergence.py` corresponding to figure 1 -* `performance_profiles.py` corresponding to figure 2 +* `convergence.py` : convergence of the four compared methods on a single problem instance +* `performance_profiles.py` : re-generate figure 1 of the "Screen & Relax" paper To run experiments, make sure to be located at project root and run the corresponding python file : ```bash @@ -57,7 +53,7 @@ python exp/ICASSP/convergence.py python exp/ICASSP/performance_profiles.py ``` -You can also edit the experimental setup directly in the script, which is well documented. +You can also edit the experimental setup directly in the script, just follow the comments documenting the code. ## Running demo examples @@ -66,7 +62,7 @@ We also provide a notebook explaining how to use the `srenet` package. You can u cd notebooks jupyter notebook ``` -You may have to install Jupyter. Please refer to the [Jupyter website](https://jupyter.org). +You may have to install Jupyter. Please refer to the [Jupyter website](https://jupyter.org) for install instructions. ## Licence