iab classification model updates

src/streamlit_app.py

@@ -93,11 +93,11 @@ def get_data_and_embeddings():
 
 
 st.set_page_config(page_title="IAB Classifier App", layout="wide")
-st.title("🧠 IAB Classifier App")
+st.title("IAB Classifier App")
 
 # Load data
 texts, prior_labels, X = get_data_and_embeddings()
-st.markdown("### 🧭 Reference sentence for similarity")
+st.markdown("### Reference sentence for similarity")
 reference = st.text_area("Type something like 'business related'")
 prediction_choice = st.checkbox("try our iab model prediction for this")
 
@@ -135,7 +135,7 @@ if reference:
     top_size = st.slider("number of similar items", 1, 100, 5)
     top_candidates = [(row["text"], row["sim"], row["label"]) for row in df_emb.to_dict(orient="records")][:top_size]
 
-    st.markdown("### 🧪 Similar example(s)")
+    st.markdown("### Similar example(s)")
     if not top_candidates:
         st.info("No more similar examples.")
     else:
@@ -149,7 +149,7 @@ if reference:
 
         # Left: What training data says
         with col1:
-            st.markdown("#### 📚 What Training Data Says")
+            st.markdown("#### What Training Data Says")
             fig1, ax1 = plt.subplots()
             top_labelled_df['label'].value_counts(normalize=True).sort_values().plot(kind='barh', ax=ax1, color="lightcoral")
             ax1.set_title("Label Distribution")
@@ -159,7 +159,7 @@ if reference:
 
         # Right: What model predicts
         with col2:
-            st.markdown("#### 🤖 Model Predictions")
+            st.markdown("#### Model Predictions")
             if len(preds) == 0 or not prediction_choice:
                 st.write("Model is unsure")
             else:
@@ -171,7 +171,7 @@ if reference:
                 st.pyplot(fig2)
 
 if prediction_choice and reference:
-    st.markdown("### 🔍 Model Explanation (Top Predicted Class)")
+    st.markdown("### Model Explanation (Top Predicted Class)")
 
     explainer = get_explainer()
     attributions = explainer(reference)
@@ -189,7 +189,7 @@ if prediction_choice and reference:
     st.pyplot(fig)
 
     # HTML Highlighted Text
-    st.markdown("#### 🔎 Highlighted Text Importance")
+    st.markdown("#### Highlighted Text Importance")
     html_output = explainer.visualize().data
 
     # Render in Streamlit