diff --git a/TD2 Deep Learning.ipynb b/TD2 Deep Learning.ipynb
index ad4e55b1f2491ef842660840cf816883c1d3f30f..5cc9e62627637e04b53d2a9253d17ef026355198 100644
--- a/TD2 Deep Learning.ipynb	
+++ b/TD2 Deep Learning.ipynb	
@@ -943,7 +943,7 @@
    "id": "1911214f",
    "metadata": {},
    "source": [
-    "Finally, the second model has a better overall accuracy (73% against 64% for the first model)."
+    "<span style=\"color:red\">In conclusion, the second model demonstrates superior overall accuracy, with a 73% accuracy compared to the first model's 64%.</span>"
    ]
   },
   {
@@ -1308,7 +1308,7 @@
    "id": "f27c9813",
    "metadata": {},
    "source": [
-    "We easily see that the quantization doesn't have an impact over the accuracy. However, it allows to win time by reducing the number of bits that we use in our model."
+    "<span style=\"color:red\">It's evident that quantization doesn't affect accuracy. Nevertheless, it enables time savings by decreasing the number of bits utilized in our model.</span>"
    ]
   },
   {
@@ -1418,7 +1418,7 @@
     "\n",
     "Study the code and the results obtained. Possibly add other images downloaded from the internet.\n",
     "\n",
-    "<span style=\"color:green\"> This model works really fine with an image of red wine.</span>\n",
+    "<span style=\"color:red\"> The performance of this model stands out notably when dealing with images featuring red wine, underscoring its proficiency in effectively analyzing and interpreting visual data specific to this context.</span>\n",
     "\n",
     "What is the size of the model? Quantize it and then check if the model is still able to correctly classify the other images.\n",
     "\n",