Update app.py

app.py

@@ -3,29 +3,27 @@ import gradio as gr
 import os
 import requests 
 
+# We get the token and the models API url 
 hf_token = os.getenv("HF_TOKEN")
 llama_7b = os.getenv("API_URL_LLAMA_7")
 llama_13b = os.getenv("API_URL_LLAMA_13")
 zephyr_7b = os.getenv("API_URL_ZEPHYR_7")
 
-#api_url_70b = os.getenv("API_URL_LLAMA_70")
-
 headers = {
     'Content-Type': 'application/json',
 }
 
-title = "Find the password 🔒"
-description = "In this game prototype, your goal is to discuss with the intercom to find the correct password"
-
-
-def predict(message, 
+"""
+Chat Function
+"""
+def chat(message, 
             chatbot, 
             system_prompt = "", 
             temperature = 0.9,
             max_new_tokens = 256,
             top_p = 0.6,
             repetition_penalty = 1.0,
-            model= api_url_70b):
+            model= llama_13b):
 
     # Write the system prompt
     if system_prompt != "":
@@ -94,7 +92,7 @@ def predict(message,
 
 
 additional_inputs=[
-    
+    model = gr.Dropdown(["llama_7b", "llama_13b", "zephyr_7b"], label="Model", info="Which model do you want to use?")
     gr.Textbox("", label="Optional system prompt"),
     gr.Slider(
         label="Temperature",
@@ -132,12 +130,15 @@ additional_inputs=[
         interactive=True,
         info="Penalize repeated tokens",
     )
-
 ]
 
+title = "Find the password 🔒"
+description = "In this game prototype, your goal is to discuss with the intercom to find the correct password"
+
+
 chatbot = gr.Chatbot(avatar_images=('user.png', 'bot2.png'),bubble_full_width = False)
 
-model = gr.Dropdown(["api_url_7b", "api_url_13b", "api_url_70b"], label="Model", info="Which model to use?")
+
 
 chat_interface = gr.ChatInterface(predict, 
                  title=title,