Update app.py

app.py

@@ -15,27 +15,27 @@ def respond(
     temperature,
     top_p,
 ):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
+    # Format the conversation in Mistral's instruction format
+    prompt = f"<s>[INST] {system_message} [/INST]\n\n"
+    
+    for user_msg, assistant_msg in history:
+        if user_msg:
+            prompt += f"[INST] {user_msg} [/INST]\n"
+        if assistant_msg:
+            prompt += f"{assistant_msg}\n\n"
+    
+    prompt += f"[INST] {message} [/INST]\n"
 
     response = ""
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
+    for token in client.text_generation(
+        prompt,
+        max_new_tokens=max_tokens,
         stream=True,
         temperature=temperature,
         top_p=top_p,
+        do_sample=True,
     ):
-        token = message.choices[0].delta.content
-
         response += token
         yield response
 
@@ -46,7 +46,7 @@ For information on how to customize the ChatInterface, peruse the gradio docs: h
 demo = gr.ChatInterface(
     respond,
     additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
+        gr.Textbox(value="You are a Microsoft 365 data management assistant.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
         gr.Slider(