Update handler.py

handler.py

@@ -7,7 +7,7 @@ class EndpointHandler:
     def __init__(self, path=""):
         # load model and processor from path
         self.processor = AutoProcessor.from_pretrained(path)
-        self.model = Blip2ForConditionalGeneration.from_pretrained(path, torch_dtype=torch.float16).to("cuda")
+        self.model = Blip2ForConditionalGeneration.from_pretrained(path, device_map="auto", load_in_8bit=True).to("cuda")
 
     def __call__(self, inputs: Dict[str, Any]) -> Dict[str, str]:
         """