Update app.py

app.py

@@ -10,7 +10,7 @@ client = OpenAI(
     base_url="https://api.runpod.ai/v2/vllm-k0g4c60zor9xuu/openai/v1",
 )
 
-def get_response(user_message, history):
+def get_response(user_message, history, verbosity):
     # Format the history for the OpenAI call
     history_openai_format = []
     for human, assistant in history:
@@ -20,12 +20,15 @@ def get_response(user_message, history):
             history_openai_format.append({"role": "assistant", "content": assistant})
     history_openai_format.append({"role": "user", "content": user_message})
 
+    # Adjust the temperature based on the verbosity level
+    temperature = 0.5 if verbosity == "Balanced" else 0.3 if verbosity == "Concise" else 0.7
+
     # Make the API call
     response = client.chat.completions.create(
         model='ambrosfitz/llama-3-history',
         messages=history_openai_format,
-        temperature=0.5,
-        max_tokens=1028
+        temperature=temperature,
+        max_tokens=150
     )
 
     # Access the text response
@@ -40,17 +43,18 @@ with gr.Blocks() as demo:
     chatbot = gr.Chatbot()
     msg = gr.Textbox()
     clear = gr.Button("Clear")
+    verbosity = gr.Radio(["Concise", "Balanced", "Detailed"], value="Balanced", label="Verbosity")
 
-    def user(user_message, history):
+    def user(user_message, history, verbosity):
         if not user_message.strip():
             return "", history
-        bot_response = get_response(user_message, history)
+        bot_response = get_response(user_message, history, verbosity)
         return "", history + [[user_message, bot_response]]
 
     def clear_chat():
         return "", []  # Clear the chat history
 
-    msg.submit(user, inputs=[msg, chatbot], outputs=[msg, chatbot])
+    msg.submit(user, inputs=[msg, chatbot, verbosity], outputs=[msg, chatbot])
     clear.click(clear_chat, inputs=None, outputs=[msg, chatbot])
 
 demo.launch()