Add history to interface

app.py

@@ -1,17 +1,21 @@
 import gradio as gr
 from transformers import pipeline
 
-# Load model locally in the Space
+# Load model locally
 pipe = pipeline("text-generation", model="HuggingFaceTB/SmolLM3-3B-Base")
 
-def respond(message, system_message, max_tokens, temperature, top_p):
-    # Combine system message + history into a single prompt
+def respond(message, system_message, max_tokens, temperature, top_p, history):
+    # Start conversation with system message
     conversation_text = system_message + "\n"
+    
+    # Append history
     for user_msg, bot_msg in history:
         if user_msg:
             conversation_text += f"User: {user_msg}\n"
         if bot_msg:
             conversation_text += f"Assistant: {bot_msg}\n"
+    
+    # Current user message
     conversation_text += f"User: {message}\nAssistant:"
 
     # Generate response
@@ -23,31 +27,29 @@ def respond(message, system_message, max_tokens, temperature, top_p):
         do_sample=True
     )
 
-    # Pipeline returns a list of dicts; take first
-    return outputs[0]["generated_text"].replace(conversation_text, "").strip()
+    assistant_reply = outputs[0]["generated_text"].replace(conversation_text, "").strip()
 
-# Interface exposes a real /run/predict endpoint
+    # Update history
+    history.append([message, assistant_reply])
+
+    return assistant_reply, history
+
+# Gradio interface
 iface = gr.Interface(
     fn=respond,
     inputs=[
         gr.Textbox(label="Message"),
         gr.Textbox(label="System message", value="You are a friendly Chatbot."),
-        gr.Slider(minimum=1, maximum=2048, value=512, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, label="Temperature"),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, label="Top-p")
+        gr.Slider(1, 2048, value=512, label="Max new tokens"),
+        gr.Slider(0.1, 4.0, value=0.7, label="Temperature"),
+        gr.Slider(0.1, 1.0, value=0.95, label="Top-p"),
+        gr.State([])  # Holds history as a list of [user, bot] tuples
     ],
-    outputs=gr.Textbox(label="Assistant Reply")
+    outputs=[
+        gr.Textbox(label="Assistant Reply"),
+        gr.State()  # Return updated history
+    ]
 )
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        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"),
-        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
-    ],
-)
-"""
+
 if __name__ == "__main__":
     iface.launch(share=True)