diff --git a/README.Rmd b/README.Rmd
index 3241630..300b26e 100644
--- a/README.Rmd
+++ b/README.Rmd
@@ -41,7 +41,7 @@ devtools::install_github("imbs-hl/fuseMLR")
The following figure illustrates the general architecture of `fuseMLR`:
```{r, echo=FALSE, out.width="70%", out.height="100%"}
-knitr::include_graphics("README_files/figure-gfm/fuseMLR_4_blue_bgr.png")
+knitr::include_graphics("README_files/figure-gfm/structure.png")
```
The general architecture of `fuseMLR` includes the storage classes `Training`, `Layer`, and `MetaLayer`. `Layer` and `MetaLayer` are stored within a `Training` instance, while `Data`, `Learner`, and `Varselect` (for variable selection) are stored within a `Layer` instance. These components can be used to automatically build and store the `Model` or `MetaModel`.
diff --git a/README.md b/README.md
index e581dba..5b57e5b 100644
--- a/README.md
+++ b/README.md
@@ -57,7 +57,7 @@ devtools::install_github("imbs-hl/fuseMLR")
The following figure illustrates the general architecture of `fuseMLR`:
-
+
The general architecture of `fuseMLR` includes the storage classes
`Training`, `Layer`, and `MetaLayer`. `Layer` and `MetaLayer` are stored
diff --git a/README_files/figure-gfm/fusemlrcheatsheet.pdf b/README_files/figure-gfm/fusemlrcheatsheet.pdf
index bf97ce9..5c80ee2 100644
Binary files a/README_files/figure-gfm/fusemlrcheatsheet.pdf and b/README_files/figure-gfm/fusemlrcheatsheet.pdf differ
diff --git a/README_files/figure-gfm/structure.pdf b/README_files/figure-gfm/structure.pdf
new file mode 100644
index 0000000..f544167
Binary files /dev/null and b/README_files/figure-gfm/structure.pdf differ
diff --git a/README_files/figure-gfm/structure.png b/README_files/figure-gfm/structure.png
new file mode 100644
index 0000000..d2cd504
Binary files /dev/null and b/README_files/figure-gfm/structure.png differ