diff --git a/hugo_blag/content/posts/2018-04-08-recommender-systems-1.md b/hugo_blag/content/posts/2018-04-08-recommender-systems-1/index.md
similarity index 99%
rename from hugo_blag/content/posts/2018-04-08-recommender-systems-1.md
rename to hugo_blag/content/posts/2018-04-08-recommender-systems-1/index.md
index e6aced1..7537ff1 100644
--- a/hugo_blag/content/posts/2018-04-08-recommender-systems-1.md
+++ b/hugo_blag/content/posts/2018-04-08-recommender-systems-1/index.md
@@ -1517,9 +1517,7 @@ for i,item in enumerate(factors[:50,:]):
 plt.show()
 {{< / highlight >}}
 
-
-![png](../images/2018-04-08-recommenders/output_94_0.png)
-
+![png](./output_94_0.png)
 
 And here are the other two:
 
@@ -1534,7 +1532,7 @@ plt.show()
 {{< / highlight >}}
 
 
-![png](../images/2018-04-08-recommenders/output_96_0.png)
+![png](./output_96_0.png)
 
 
 Below is another way of visualizing. Neither the code nor the result are very pretty, but it divides the entire latent space into a 2D grid, identifies the top few movies (ranked by number of ratings) in each grid square, and prints the resultant grid.
diff --git a/images/2018-04-08-recommenders/output_94_0.png b/hugo_blag/content/posts/2018-04-08-recommender-systems-1/output_94_0.png
similarity index 100%
rename from images/2018-04-08-recommenders/output_94_0.png
rename to hugo_blag/content/posts/2018-04-08-recommender-systems-1/output_94_0.png
diff --git a/images/2018-04-08-recommenders/output_96_0.png b/hugo_blag/content/posts/2018-04-08-recommender-systems-1/output_96_0.png
similarity index 100%
rename from images/2018-04-08-recommenders/output_96_0.png
rename to hugo_blag/content/posts/2018-04-08-recommender-systems-1/output_96_0.png