update documentation

app.py

@@ -66,6 +66,7 @@ article = """
 \n
 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 \n
+# Context
 I fine tuned Open Ai's CLIP model on both text (tweets) and images of emoji's!\n
 The current model you can play with is fine-tuned on 15 samples per emoji.
 
@@ -73,17 +74,20 @@ The current model you can play with is fine-tuned on 15 samples per emoji.
 - dataset: https://huggingface.co/datasets/vincentclaes/emoji-predictor \n
 - profile: https://huggingface.co/vincentclaes \n
 
-- Below you can find a table with the precision for predictions and suggestions\n 
+# Performance
+
+Below you can find a table with the precision for predictions and suggestions 
 for a range of samples per emoji we fine-tuned CLIP on.
 
+### Prediction vs. Suggestion
 - The column "Prediction" indicates the precision for predicting the right emoji.
 
-- Since there can be some confusion about the right emoji for a tweet, \n
-I also tried to present 4 suggestions. If 1 of the 4 suggestions is the same as the label, \n
+- Since there can be some confusion about the right emoji for a tweet,
+I also tried to present 4 suggestions. If 1 of the 4 suggestions is the same as the label,
 I consider it a valid prediction. See the column "Suggestion".
 
-- Randomly predicting an emoji would have a precision of 1/32 or 0.0325.\n
-Randomly suggesting an emoji would have a precision of 4/32 or 0.12.
+- Randomly predicting an emoji would have a precision of 1/32 or 0.0325.
+- Randomly suggesting an emoji would have a precision of 4/32 or 0.12.
 
 
            | Samples  | Prediction  | Suggestion  |