Update app.py

app.py

@@ -8,7 +8,7 @@ zero = torch.Tensor([0]).cuda()
 print(zero.device) # <-- 'cpu' 🤔
 
 @spaces.GPU
-def greet(n):
+def greet(llm):
     print(zero.device) # <-- 'cuda:0' 🤗
     grammar = LlamaGrammar.from_string('''
     root ::= sentence
@@ -66,6 +66,6 @@ def load_model(fp):
     return llm
 
 llm = load_model(download_model())
-demo = gr.Interface(fn=greet, inputs=gr.Number(), outputs=gr.Text())
+demo = gr.Interface(fn=greet, inputs=llm, outputs=gr.Text())
 demo.launch(share=False)