Improved inference

app.py

@@ -190,8 +190,8 @@ if prompt := st.chat_input("Chat with Ikigai Docs..."):
         placeholder = st.empty()
         if st.session_state.rag_enabled :
             for chunk in chat_stream :
-                if chunk.token['text']!='</s>' :
-                    full_response += chunk.token["text"]
+                if chunk.token.text!='</s>' :
+                    full_response += chunk.token.text
 
                 placeholder.markdown(full_response + " |")
             
@@ -200,7 +200,7 @@ if prompt := st.chat_input("Chat with Ikigai Docs..."):
         else :
             for chunk in chat_stream :
                 if chunk.token['text']!='</s>' :
-                    full_response += chunk.token["text"]
+                    full_response += chunk.token.text
                 placeholder.markdown(full_response + " |")
             placeholder.markdown(full_response)