Update app.py

app.py

@@ -1,20 +1,22 @@
+import os
 import gradio as gr
 from huggingface_hub import InferenceClient
 
+SYSTEM_PROMPT = os.environ["SYSTEM_PROMPT"]
 client = InferenceClient("HuggingFaceTB/SmolLM2-1.7B-Instruct")
 
 def chat(message, history):
-    history.append({"role": "user", "content": message})
-    stream = client.chat.completions.create(
-        messages=history,
+    messages = [{"role": "system", "content": SYSTEM_PROMPT}] + history
+    messages.append({"role": "user", "content": message})
+    
+    result = ""
+    for chunk in client.chat.completions.create(
+        messages=messages,
         max_tokens=1024,
         stream=True
-    )
-    result = ""
-    for chunk in stream:
+    ):
         result += chunk.choices[0].delta.content
         yield result
-    yield result
 
 with gr.Blocks(
     title="🤏 SmolLM Instruct",
@@ -30,7 +32,7 @@ with gr.Blocks(
     gr.ChatInterface(
         fn=chat,
         type="messages",
-        examples=["Hi!"]
+        examples=["Hi!", "Who are you?"]
     )
 
     with gr.Accordion("ℹ️ About", open=False):