Adjusted the model for concise responses

app.py

@@ -3,9 +3,12 @@ from huggingface_hub import InferenceClient
 
 # ---------------- CONFIG ----------------
 MODEL_REPO = "HuggingFaceH4/zephyr-7b-beta"
-SYSTEM_PROMPT_DEFAULT = "You are Zephyr, a helpful, concise and polite AI assistant."
+SYSTEM_PROMPT_DEFAULT = (
+    "You are Zephyr, a concise and polite AI assistant. "
+    "Answer briefly unless the user specifically asks for detail."
+)
 
-MAX_NEW_TOKENS_DEFAULT = 512
+MAX_NEW_TOKENS_DEFAULT = 128
 TEMP_DEFAULT = 0.7
 TOP_P_DEFAULT = 0.95
 
@@ -13,14 +16,15 @@ TOP_P_DEFAULT = 0.95
 client = InferenceClient(MODEL_REPO)
 
 # ---------------- CHAT FUNCTION ----------------
-def stream_response(message, chat_history, system_message, max_tokens, temperature, top_p):
+def stream_response(message, chat_history, system_message, max_tokens, temperature, top_p, response_style):
+    if response_style == "concise":
+        system_message += " Keep answers short and direct."
+    elif response_style == "detailed":
+        system_message += " Provide more explanation and context when helpful."
+    elif response_style == "essay":
+        system_message += " Write long, structured, essay-style responses."
+
     messages = [{"role": "system", "content": system_message}]
-    for user_msg, bot_msg in chat_history:
-        if user_msg:
-            messages.append({"role": "user", "content": user_msg})
-        if bot_msg:
-            messages.append({"role": "assistant", "content": bot_msg})
-    messages.append({"role": "user", "content": message})
 
     response = ""
     for msg in client.chat_completion(
@@ -73,14 +77,14 @@ with gr.Blocks(theme=gr.themes.Soft(primary_hue="violet", secondary_hue="pink"))
 
     # Events (streaming response)
     send_btn.click(
-        stream_response,
-        [msg, chatbot, system_prompt, max_tokens, temperature, top_p],
-        [msg, chatbot]
+    stream_response,
+    [msg, chatbot, system_prompt, max_tokens, temperature, top_p, response_style],
+    [msg, chatbot]
     )
     msg.submit(
-        stream_response,
-        [msg, chatbot, system_prompt, max_tokens, temperature, top_p],
-        [msg, chatbot]
+    stream_response,
+    [msg, chatbot, system_prompt, max_tokens, temperature, top_p, response_style],
+    [msg, chatbot]
     )
     clear_btn.click(lambda: None, None, chatbot, queue=False)