fix: dependency

README.md

@@ -50,6 +50,7 @@ This library allows you to look for certain inefficiencies withing your convolut
 ever training the model.
 You can do this simply by importing your architecture into the format of RFA-Toolbox and then use the in-build functions
 to visualize your architecture using GraphViz.
+The visualization will automatically mark layers predicted to be unproductive red and critical layers, that are potentially unproductive orange.
 This is especially useful if you plan to train your model on resolutions that are substantially lower than the
 design-resolution of most models.
 As an alternative, you can also use the graph from RFA-Toolbox to hook RFA-toolbox more directly into your program.