Update app.py

app.py

@@ -3,16 +3,21 @@ import gradio as gr
 
 client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
 
-def format_prompt(message, history):
+def format_prompt(preprompt, message, history):
   prompt = "<s>"
+  
+  if preprompt:
+    prompt += f"[INST] {preprompt} [/INST]"
+
   for user_prompt, bot_response in history:
     prompt += f"[INST] {user_prompt} [/INST]"
     prompt += f" {bot_response}</s> "
   prompt += f"[INST] {message} [/INST]"
+  
   return prompt
 
 def generate(
-    prompt, history, temperature=0.2, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,
+    prompt, history, temperature=0.2, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0, preprompt=None
 ):
     temperature = float(temperature)
     if temperature < 1e-2:
@@ -28,7 +33,7 @@ def generate(
         seed=42,
     )
 
-    formatted_prompt = format_prompt(prompt, history)
+    formatted_prompt = format_prompt(preprompt, prompt, history)
 
     stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
     output = ""
@@ -40,7 +45,7 @@ def generate(
 
     
 mychatbot = gr.Chatbot(
-    avatar_images=["./user.png", "./botm.png"], bubble_full_width=False, show_label=False, show_copy_button=True, likeable=True,)
+    avatar_images=["./berger.jpg", "./tavernier.jpg"], bubble_full_width=False, show_label=False, show_copy_button=True, likeable=True,)
 
 demo = gr.ChatInterface(fn=generate, 
                         chatbot=mychatbot,
@@ -49,4 +54,4 @@ demo = gr.ChatInterface(fn=generate,
                         undo_btn=None
                        )
 
-demo.queue().launch(show_api=True)
+demo.serve(preprompt="Bienvenue dans ma taverne!").launch(show_api=True)