Update app.py

app.py

@@ -19,13 +19,13 @@ def init_model():
         nn.Sigmoid()
     )
     
-    model_path = 'model.pt'
+    model_path = "model.pt"
     model.load_state_dict(torch.load(model_path, map_location=torch.device("cpu")))
     model.eval()
     return model
 
-cats = ['Computer Science', 'Economics', 'Electrical Engineering', 
-        'Mathematics', 'Physics', 'Biology', 'Finance', 'Statistics']
+cats = ["Computer Science", "Economics", "Electrical Engineering", 
+        "Mathematics", "Physics", "Biology", "Finance", "Statistics"]
 
 def predict(outputs):
     top = 0
@@ -34,14 +34,19 @@ def predict(outputs):
     top_cats = []
     top_probs = []
 
+    first = True
     for prob, cat in sorted(zip(probs, cats), reverse=True):
+        if first:
+            if cat == "Computer Science":
+                st.write("Today everything is connected with Computer Science"
+            first = False
         if top < 95:
             percent = prob * 100
             top += percent
             top_cats.append(cat)
             top_probs.append(round(percent, 1))
             
-    chart_data = pd.DataFrame(top_probs, index=top_cats, columns=['percent'])
+    chart_data = pd.DataFrame(top_probs, index=top_cats, columns=["percent"])
     st.bar_chart(chart_data)
 
 tokenizer = RobertaTokenizer.from_pretrained("roberta-large-mnli")
@@ -59,7 +64,7 @@ if not title:
     st.warning("Please fill in required fields")
 else:    
     st.markdown("### Result")
-    encoded_input = tokenizer(title + '. ' + abstract, return_tensors='pt', padding=True, 
+    encoded_input = tokenizer(title + ". " + abstract, return_tensors="pt", padding=True, 
                           max_length = 512, truncation=True)
     with torch.no_grad():
         outputs = model(**encoded_input).pooler_output[:, 0, :]