diff --git a/README.md b/README.md
index 259aac4b5531b8ab7336e4ce6702e2d0009281ea..d03f69edb5d19f199001478049b90aa00de6617a 100644
--- a/README.md
+++ b/README.md
@@ -1,3 +1,9 @@
+Please find in this file the codes used for the second part of the third TD. 
+I did the first one using google colab but not for the second one. You can thus find the codes I used bellow. 
+Thank you for your understanding. 
+Pierre Danjou 
+
+
 To adapt the code to apply the ViT model on CIFAR dataset : 
 
 ```python
@@ -232,4 +238,7 @@ print(
         np.sum(class_total),
     )
 )
-```
\ No newline at end of file
+```
+
+
+After all the test, we can find an accuracy for this model of around 42% which is better than the KNN but lower than the model used in TD2. 
\ No newline at end of file