Update README.md

xshhhm · Mar 4, 2017 · a780477 · a780477
1 parent b4bcabd
commit a780477
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/README.md b/README.md
@@ -104,8 +104,8 @@ We observe that in the earlier experiment we cannot learn above markovity close
 
 For a 16-size 2 layer network, with sequence length of 8, we were able to train for markovity 10 very well \(thus even though we do not explicitly backproagate, there is still some learning\).
 
-![train-1](images/loss_8.png\)
-![val-1]\(images/val_loss_8.png)
+![train-1](images/loss_8.png)
+![val-1](images/val_loss_8.png)
 
 #### Training very large networks