do cuda

handler.py

@@ -13,12 +13,12 @@ class EndpointHandler():
         #self.model = MarkupLMModel.from_pretrained("jasper-lu/pix2struct_embedding")
         self.processor = AutoProcessor.from_pretrained(MODEL)
         self.processor.image_processor.is_vqa = False
-        self.model = Pix2StructVisionModel.from_pretrained(MODEL)
+        self.model = Pix2StructVisionModel.from_pretrained(MODEL).cuda()
 
     def __call__(self, data: Any) -> List[List[Dict[str, float]]]:
         url = data.pop("inputs", data)
         image = Image.open(requests.get(url, stream=True).raw)
-        inputs = self.processor(images=image, return_tensors="pt")
+        inputs = self.processor(images=image, return_tensors="pt").cuda()
 
         with torch.no_grad():
             outputs = self.model(**inputs)