Implemented Fixes

app.py

@@ -16,17 +16,17 @@ def respond(
     temperature,
     top_p,
 ):
-    # Format the chat history for the DialoGPT model
-    full_conversation = ""
+    # Format the chat history for the DialoGPT model using the 'messages' format
+    conversation = [{"role": "system", "content": system_message}]
     for user_msg, bot_msg in history:
         if user_msg:
-            full_conversation += f"User: {user_msg}\n"
+            conversation.append({"role": "user", "content": user_msg})
         if bot_msg:
-            full_conversation += f"DialoGPT: {bot_msg}\n"
-    full_conversation += f"User: {message}\nDialoGPT:"
+            conversation.append({"role": "assistant", "content": bot_msg})
+    conversation.append({"role": "user", "content": message})
 
     # Tokenize input and generate response
-    inputs = tokenizer.encode(full_conversation, return_tensors="pt")
+    inputs = tokenizer.encode(" ".join([msg["content"] for msg in conversation]), return_tensors="pt")
     outputs = model.generate(
         inputs,
         max_length=max_tokens,
@@ -38,9 +38,11 @@ def respond(
     return response
 
 # Gradio Chat Interface
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
+demo = gr.Interface(
+    fn=respond,
+    inputs=[
+        gr.Textbox(label="Message"),
+        gr.State(),  # For history
         gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
@@ -52,9 +54,8 @@ demo = gr.ChatInterface(
             label="Top-p (nucleus sampling)",
         ),
     ],
+    outputs="text",
 )
 
 if __name__ == "__main__":
-    # Launch the Gradio app without 'enable_api'
     demo.launch()
-