try

app.py

@@ -59,7 +59,7 @@ def get_top_chunks(query):
     
 client = InferenceClient(
     model='Qwen/Qwen2.5-72B-Instruct'
-    #token = os.getenv('HF_TOKEN')
+    token = 'HF_TOKEN'
 ) 
 
 #client is where you can change the LLM model!
@@ -117,4 +117,4 @@ with gr.Blocks(theme='hmb/amethyst') as demo:
     send.click(respond, [msg, chatbot], [msg, chatbot])
 
 demo.launch()
-chatbot.launch()
+