Update app.py

app.py

@@ -3,28 +3,30 @@ from huggingface_hub import InferenceClient
 
 client = InferenceClient("Futuresony/future_ai_12_10_2024.gguf")
 
-# Store chat history
-def format_alpaca_prompt(user_input, history, system_prompt):
-    """Formats input in Alpaca/LLaMA style with history"""
-    history_text = "\n".join([f"User: {h[0]}\nAssistant: {h[1]}" for h in history])
-    prompt = f"""{system_prompt}\n{history_text}\nUser: {user_input}\nAssistant:"""
+# Store conversation history
+def format_alpaca_prompt(history, user_input, system_prompt):
+    """Formats input in Alpaca/LLaMA style with conversation history"""
+    formatted_history = "\n".join([f"User: {h[0]}\nAssistant: {h[1]}" for h in history])
+    prompt = f"""{system_prompt}\n{formatted_history}\nUser: {user_input}\nAssistant:"""
     return prompt
 
 def respond(message, history, system_message, max_tokens, temperature, top_p):
-    formatted_prompt = format_alpaca_prompt(message, history, system_message)
-
+    formatted_prompt = format_alpaca_prompt(history, message, system_message)
+    
     response = client.text_generation(
         formatted_prompt,
         max_new_tokens=max_tokens,
         temperature=temperature,
         top_p=top_p,
     )
-
-    # ✅ Extract only the response
+    
+    # Extract only the response
     cleaned_response = response.strip()
     
-    history.append((message, cleaned_response))  # ✅ Store conversation history
-    yield cleaned_response  # ✅ Output only the answer
+    # Update history
+    history.append((message, cleaned_response))
+    
+    yield cleaned_response  # Output only the answer
 
 demo = gr.ChatInterface(
     respond,