compile, and remove input from output

app.py

@@ -5,11 +5,19 @@ import torch
 model_name = "AddieFoote0/arithmetic-300M-reference"
 model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.bfloat16)
 tokenizer = AutoTokenizer.from_pretrained(model_name)
+if hasattr(torch, "compile"):
+    model = torch.compile(model)
+    print("compiled model")
+else:
+    print("no compile")
+
 
 def generate_response(prompt):
     inputs = tokenizer(prompt, return_tensors="pt")
-    outputs = model.generate(**inputs, max_new_tokens=10, temperature=1.0)
-    return tokenizer.decode(outputs[0], skip_special_tokens=True)
+    outputs = model.generate(**inputs, max_new_tokens=5, temperature=1.0)
+    full_output = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    new_tokens = full_output[len(prompt):]
+    return new_tokens
 
 iface = gr.Interface(
     fn=generate_response,