added yield response

app.py

@@ -121,7 +121,7 @@ cleaned_chunks = preprocess_text(poverty_and_education)
 chunk_embeddings = create_embeddings(cleaned_chunks)
 #AI API being used
 client= InferenceClient("Qwen/Qwen2.5-7B-Instruct-1M")
-
+response=""
 #defining role of AI and user
 def respond(message,history):
     information = get_top_chunks(message, chunk_embeddings, cleaned_chunks)
@@ -134,9 +134,11 @@ def respond(message,history):
 
     messages.append({"role":"user", "content": message})
 
-    response=client.chat_completion(messages, max_tokens=100) #capping how many words the LLM is allowed to generate as a respond (100 words)
-
-    return response['choices'][0]['message']['content'].strip() #storing value of response in a readable format to display
+    response=client.chat_completion(messages, stream=True, max_tokens=100) #capping how many words the LLM is allowed to generate as a respond (100 words)
+    for message in client.chat_completion(messages):
+        token = message.choices[0].delta.content
+        response+=token
+    yield response['choices'][0]['message']['content'].strip() #storing value of response in a readable format to display
 
 ### STEP 6
 # Call the preprocess_text function and store the result in a cleaned_chunks variable