Update inference.py

inference.py

@@ -18,6 +18,7 @@ class PaliGemmaModel:
         self.model = PaliGemmaForConditionalGeneration.from_pretrained(self.model_id).eval().to(self.device)
         self.processor = PaliGemmaProcessor.from_pretrained(self.model_id)
 
+    @spaces.GPU
     def infer(self, image: PIL.Image.Image, text: str, max_new_tokens: int) -> str:
         inputs = self.processor(text=text, images=image, return_tensors="pt")
         inputs = {k: v.to(self.device) for k, v in inputs.items()}  # Move inputs to the correct device