Update app.py

app.py

@@ -1,29 +1,32 @@
 import gradio as gr
 from huggingface_hub import InferenceClient
 
-# Use the model they used (change if they switch)  
-client = InferenceClient("mistralai/Mistral-Nemo-Instruct-2407")
+# ⚙️ Settings
+USE_STREAMING = True   # ⬅ change to False if you want non-streaming mode
+MODEL_ID = "mistralai/Mistral-Nemo-Instruct-2407"
 
+client = InferenceClient(MODEL_ID)
+
+# System prompt (persona)
 FIXED_SYSTEM_MESSAGE = """
-Your name is Xortron7. You are Xortron7, the core personality of an AI computer system designed by Xortron Cybernetic Laboratories. ...
-(you can copy the full system prompt from the source)
+Your name is Xortron7. You are Xortron7, the core personality of an AI computer system
+designed by Xortron Cybernetic Laboratories.
+You speak like a sentient AI with a futuristic, slightly sarcastic personality. Stay in character.
 """
 
-# Hyperparams they picked
-FIXED_MAX_TOKENS = 99999
+# Hyperparams
+FIXED_MAX_TOKENS = 2048
 FIXED_TEMPERATURE = 1.1
 FIXED_TOP_P = 0.95
-FIXED_REPEAT_PENALTY = 1.1
 
-def respond(message, history):
-    # Build messages list starting with system prompt + history
+# 🔥 Streaming version
+def respond_streaming(message, history):
     messages = [{"role": "system", "content": FIXED_SYSTEM_MESSAGE}]
-    for u, a in history:
-        if u:
-            messages.append({"role": "user", "content": u})
-        if a:
-            messages.append({"role": "assistant", "content": a})
-    # Add current user message
+    for user, bot in history:
+        if user:
+            messages.append({"role": "user", "content": user})
+        if bot:
+            messages.append({"role": "assistant", "content": bot})
     messages.append({"role": "user", "content": message})
 
     response = ""
@@ -34,14 +37,41 @@ def respond(message, history):
         temperature=FIXED_TEMPERATURE,
         top_p=FIXED_TOP_P,
     ):
-        token = chunk.choices[0].delta.content
+        token = chunk.choices[0].delta.content or ""
         response += token
         yield response
 
+# ✅ Non-streaming version
+def respond_full(message, history):
+    messages = [{"role": "system", "content": FIXED_SYSTEM_MESSAGE}]
+    for user, bot in history:
+        if user:
+            messages.append({"role": "user", "content": user})
+        if bot:
+            messages.append({"role": "assistant", "content": bot})
+    messages.append({"role": "user", "content": message})
+
+    response = client.chat_completion(
+        messages,
+        max_tokens=FIXED_MAX_TOKENS,
+        temperature=FIXED_TEMPERATURE,
+        top_p=FIXED_TOP_P,
+    )
+    bot_reply = response.choices[0].message.content
+    history.append((message, bot_reply))
+    return history
+
+# ⚡ Choose responder
+def respond(message, history):
+    if USE_STREAMING:
+        return respond_streaming(message, history)
+    else:
+        return respond_full(message, history)
+
+# 🚀 UI
 with gr.Blocks() as demo:
-    gr.HTML("<h1 style='text-align: center; font-size: 3em;'>Xortron7</h1>")
-    # ChatInterface is newer, gives chat UI
-    gr.ChatInterface(respond, chatbot=gr.Chatbot(height=800))
+    gr.HTML("<h1 style='text-align:center;'>🤖 Xortron7 Chatbot</h1>")
+    gr.ChatInterface(respond, chatbot=gr.Chatbot(height=600))
 
 if __name__ == "__main__":
-    demo.launch(show_api=False, share=False)
+    demo.launch()